Desmosomal connectomics of all somatic muscles in an annelid larva

  1. Sanja Jasek
  2. Csaba Verasztó
  3. Emelie Brodrick
  4. Réza Shahidi
  5. Tom Kazimiers
  6. Alexandra Kerbl
  7. Gáspár Jékely  Is a corresponding author
  1. Living Systems Institute, University of Exeter, United Kingdom
  2. Janelia Research Campus, United States
  3. kazmos GmbH, Germany

Abstract

Cells form networks in animal tissues through synaptic, chemical, and adhesive links. Invertebrate muscle cells often connect to other cells through desmosomes, adhesive junctions anchored by intermediate filaments. To study desmosomal networks, we skeletonised 853 muscle cells and their desmosomal partners in volume electron microscopy data covering an entire larva of the annelid Platynereis. Muscle cells adhere to each other, to epithelial, glial, ciliated, and bristle-producing cells and to the basal lamina, forming a desmosomal connectome of over 2000 cells. The aciculae – chitin rods that form an endoskeleton in the segmental appendages – are highly connected hubs in this network. This agrees with the many degrees of freedom of their movement, as revealed by video microscopy. Mapping motoneuron synapses to the desmosomal connectome allowed us to infer the extent of tissue influenced by motoneurons. Our work shows how cellular-level maps of synaptic and adherent force networks can elucidate body mechanics.

Editor's evaluation

This paper is based on the digital reconstruction of a serial EM stack of a larva of the annelid Platynereis and presents a complete 3D map of all desmosomes between somatic muscle cells and their attachment partners. This resource is of interest to scientists in several fields: motor control, high-resolution anatomy, and network analyses. With the first comprehensive and complete mapping of muscle-to-body connectivity through desmosomes in an annelid larva, it has the potential to close a missing link and make progress towards understanding in a "holistic" way how a complex neural circuitry controls an equally complex pattern of movement/behavior.

https://doi.org/10.7554/eLife.71231.sa0

Introduction

Animals use somatic muscles to rapidly change body shape. Whole-body shape changes can drive swimming, crawling, or other forms of movement. Moving appendages with exo- or endoskeletons and joints can have many degrees of freedom to support walking, flying, and several other functions. This requires a large diversity of muscles with differential nervous control (Brierley et al., 2012; Charles et al., 2016; Cruce, 1974; Landmesser, 1978; McKellar et al., 2020).

To exert force and induce movements, contracting muscles must attach to other body parts. In many animals, muscle–force coupling is provided by desmosomes – stable adhesive junctions anchored by intracellular intermediate filaments. Desmosomes occur in cardiac and somatic muscles of both vertebrates and invertebrates. In the zebrafish heart for example, desmosomes connect cardiac myocytes to each other (Lafontant et al., 2013). In the lamprey, desmosomes frequently unite adjacent muscle fibres for lateral force transmission (Nakao, 1975). In the nematode Caenorhabditis elegans, hemidesmosomes link the pharyngeal muscles to the basement membrane (Albertson and Thomson, 1976) and the body-wall muscles to hypodermal cells that lie underneath the cuticle (Francis and Waterston, 1991). In leech, muscle cells have a high density of hemidesmosomes anchoring the cells to the basal lamina (Pumplin and Muller, 1983). In polychaete annelids, hemidesmosomes and desmosomes are common on the extracellular matrix connecting muscles and epithelial cells (Purschke, 1985). In Echiura, hemidesmosomes connect the chaetal follicle cells to the extracellular matrix. Opposite these sites, muscle cells have dense plaques also linking them to the extracellular matrix (Tilic et al., 2015).

These desmosomal connections form a body-wide network through which tensile forces propagate. This network comprises all muscles and all their cellular and non-cellular (e.g. basal lamina) partners. We refer to this network as the desmosomal connectome, with muscles, other cells and basal lamina chunks as nodes and hemidesmosomes and desmosomes as links. How such networks are organised on the scale of the whole body in animals with complex muscles and appendages is not known. While whole-body neuronal connectomes (where links are synaptic connections) exist (Cook et al., 2019; Ryan et al., 2016; Verasztó et al., 2020; White et al., 1986), to our knowledge, no whole-body desmosomal connectome (where links are desmosomes) has yet been described. Unlike synapses, desmosomal connections are not directional and do not represent signal flow. Rather they indicate physical coupling where the movement of one cell will result in the movement of the connected structure.

To study the organisation of a body-wide tensile network and its motor control, here we reconstruct the desmosomal connectome of larval Platynereis dumerilii from serial electron microscopy images (volume EM) of an entire larva (Randel et al., 2015). P. dumerilii or ‘the nereid’ is a marine annelid worm that is increasingly used as a laboratory animal (Özpolat et al., 2021).

In Platynereis, muscle development starts in the planktonic, ciliated trochophore larval stage (1–2 days old) (Fischer et al., 2010). The older (around 3–6 days) nectochaete larvae have three main trunk segments, each with a pair of appendages called parapodia. The parapodia are composed of a ventral lobe (neuropodium) and a dorsal lobe (notopodium) and each lobe contains a single acicula and a bundle of chitin bristles (chaetae) (Hausen, 2005). Platynereis larvae have an additional cryptic segment between the head and the main trunk segments (Steinmetz et al., 2011). This cryptic segment – also referred to as segment 0 – lacks parapodia and gives rise to the anterior pair of tentacular cirri that start to grow around 3 days of development.

Nectochaete larvae can either swim with cilia (Jékely et al., 2008) or crawl on the substrate by muscles (Conzelmann et al., 2013; Lauri et al., 2014). Trunk muscles also support turning in swimming larvae during visual phototaxis (Randel et al., 2014) and mediate a startle response characterised by whole-body contraction and the elevation of the parapodia (Bezares-Calderón et al., 2018). Older (>6-day-old) feeding juvenile stages are mainly crawling and develop a muscular, eversible proboscis (Fischer et al., 2010), a moving gut with visceral muscles (Brunet et al., 2016; Williams et al., 2015), and several sensory appendages (palps, antennae, and cirri).

Platynereis larvae (1- to 6-day-old stage) are 150–300 µm long and thus amenable to whole-body volume-EM analysis. Here, we use an available volume-EM dataset of a 3-day-old nectochaete larva (Randel et al., 2015) to reconstruct the desmosomal connectome and analyse how motoneurons innervate muscles.

Results

Serial EM reconstruction of all muscle cells in the Platynereis larva

To analyse the organisation of the somatic musculature, we used a previously reported whole-body serial TEM dataset of a 3-day-old P. dumerilii larva (Randel et al., 2015). First, we reconstructed all differentiated muscle cells by tracing along their longitudinal axis (skeletonisation) and marking the position of their soma (Figure 1, Figure 2). We also skeletonised all other cells in the larva, including neuronal and non-neuronal cells (see below and Verasztó et al., 2020).

Figure 1 with 3 supplements see all
Whole-body desmosomal connectomics in the Platynereis larva.

(A) Stylised scanning electron micrograph of a 3-day-old Platynereis larva with the main body regions indicated. (B) Skeleton representation of the aciculae and all muscle cells in the larva, coloured by body segment. The position and size of the cell body of the aciculoblasts are shown as grey spheres. The developing tentacular cirri of the cryptic segment (segment 0) are also shown in grey for reference. Nucleus positions are not shown for the muscle cells. (C) Position of all annotated desmosomes and hemidesmosomes in the volume. Desmosomes and hemidesmosomes were annotated every 50 layers in the first round of annotations, hence the appearance of lines. (D) Skeletonised representations of all the cells that connect to muscle cells through desmosomes. Spheres represent position and size of cell nuclei. (E) Transmission electron micrographs showing examples of desmosomal muscle-attachment sites (indicated with asterisks) between muscles and an aciculoblast (left panel) or an epidermal cell (right panel). (F) Overview of the desmosomal connectome, comprising muscle cells and their desmosomal partners. Nodes represent single cells or basal lamina fragments (red, muscles; cyan, other cells; grey, basal lamina fragments), connections represent desmosomes and hemidesmosomes. Node size is proportional to the weighted degree of nodes. Edge thickness is proportional to the number of desmosomes. The graph is undirected. Panels A–D show ventral views. In B–D, the body outline and yolk outline are shown in grey. Abbreviations: sg0–3, segments 0–3. This figure can be reproduced by running the code/Figure1.R script from the Jasek_et_al GitHub repository (Jasek, 2022).

Classification and distribution of muscle cell types in the 3-day-old Platynereis larva.

(A) Morphological rendering of traced skeletons and soma position (spheres) of all acicular muscles. The table shows the cell-type classification of acicular muscles. (B) All anterior oblique muscles and their classification. (C) All posterior oblique muscles and their classification. (D) All chaetal muscles and their classification. (E) All transverse muscles (one type only). (F) All longitudinal muscles and their classification. (G) Muscles of the developing digestive system and the head and their classification. (H) All muscles, coloured by category. (I) Segmental and left–right distribution of all muscle cell types. Numbers indicate the number of cells per type and body region. The short names of the individual muscle types are based on composites of anatomical abbreviations (cf. panels A–G). Not, notopodial; neu, neuropodial; A, anterior; P, posterior; D, dorsal; V, ventral; ac, acicular; chae, chaetal; ob, oblique; trans, transverse; long, longitudinal. Full muscle names and their abbreviations are listed in Table 1. In A–H, the left panels are ventral views, the right panels lateral-left views. The body outline is shown in grey, aciculae are shown for segmental reference. This figure can be reproduced by running the code/Figure2.R script (Jasek, 2022).

Figure 2—source data 1

Source data for panel I listing the segmental and left–right distribution of all muscle cell types.

https://cdn.elifesciences.org/articles/71231/elife-71231-fig2-data1-v2.csv
Figure 2—source data 2

All muscle skeletons as point clouds with tangent vectors (dotprop).

An R data source file (.RDS).

https://cdn.elifesciences.org/articles/71231/elife-71231-fig2-data2-v2.zip

We identified 853 mononucleated muscle cells in the whole-body volume. Most of them have an obliquely striated ultrastructure, with the exception of developing muscle cells (which contain few myosin filaments), and the two MUSmed_head cells which have disorganised myosin filaments. The majority of muscle cells in the 3-day-old larva are somatic muscles (Table 1). Most visceral muscles, including gut muscles, develop at a later stage (Brunet et al., 2016). Based on their position and shape, we classified all muscle cells into eight main groups (Figure 2, Table 1): longitudinal, anterior and posterior oblique, chaetal, acicular, transverse, digestive system (developing), and head muscles. All anatomical reconstructions are available at https://catmaid.jekelylab.ex.ac.uk or can be accessed by the R code provided. For the 853 muscle cells, we also provide the skeletons as point clouds with tangent vectors (dotprops) (Figure 2—source data 2).

Table 1
Muscle cell types.

Hierarchical classification and nomenclature of all muscle cells in the 3-day-old Platynereis larva.

Anatomical name and CATMAID annotationExample CATMAID neuron name (example)Number of cells (per segment [sg] and body side)Desmosomal connections toSynaptic connections from (>3 synapses)
Acicular muscle (MUSac)NotopodialAnterior notopodial acicular muscleMUSac-notAsg1: 2 l, 2 r sg2: 2 l, 2 r sg3: 2 l, 2 rProximal base of notopodial acicula; notopodial ECsMNring, MNcrab, MNwave, MNantacic
Posterior notopodial acicular muscleMUSac-notPsg1: 2 l, 2 r sg2: 2 l, 2 r sg3: 2 l, 2 rProximal base of notopodial acicula; septal ECs and dorsal paratrochMNacic, MNwave, MNbow,
MNspider-ant, MNcrab, MNperif
Middle notopodial acicular muscleMUSac-notMsg1: 2 l, 2 r sg2: 2 l, 2 r sg3: 2 l, 2 rProximal base of notopodial acicula; parapodial ECsMNwave, MNacic, MNantaci, MNbow
Unpaired notopodial dorsal acicular muscleMUSac-notDosg1: 0 l, 0 r sg2: 1 l, 0 r sg3: 0 l, 0 rProximal base of notopodial acicula; dorsal ECsNone
NeuropodialAnterior ventral neuropodial acicular muscleMUSac-neuAVsg1: 2 l, 2 r sg2: 3 l, 3 r sg3: 3 l, 3 rProximal base of neuropodial acicula; epidermal cells near ventrolateral musclesMNcrab, MNring,
MNchae
Posterior dorsal neuropodial acicular muscleMUSac-neuPDsg1: 3 l, 3 r sg2: 3 l, 3 r sg3: 3 l, 3 rProximal base of neuropodial acicula; dorsal ECs and notopodial EC chaeFC and distal notopodial acicula (1) and dorsal paratroch (2)MNarm, sparse various others
Posterior ventral neuropodial acicular muscleMUSac-neuPVsg1: 2 l, 2 r sg2: 3 l, 3 r sg3: 2 l, 2 rProximal base of neuropodial acicula; ventral paratroch and ECs around itMNcrab, MNacicX,
MNperifac, MNspider-post, MNspider-ant, MNpostv
Neuropodial YMUSac-neuDysg1: 0 l, 0 r sg2: 1 l, 1 r sg3: 1 l, 1 rProximal and mid neuropodial acFC; neuropodial chaeFC-hemi; mid-parapodial septal ECFragments
Dorsal neuropodial muscle to notopodiumMUSac-neuDxsg1: 0 l, 0 r sg2: 2 l, 2 r sg3: 2 l, 2 rProximal neuropodial acFC; mid-distal notopodial acFC; and mid-dorsal parapodial ECs (2)MNchae, MNcrab, MNarm
Dorsal neuropodial chaetal muscleMUSac-neuDachsg1: 3 l, 3 r sg2: 3 l, 3 r sg3: 3 l, 3 rEntire length of neuropodial acicula (all acFC); mid-parapodial ECsMNarm, MNac, MNchae, MNantacic
Chaetal sac retractorMUSac-neuresg1: 1 l, 1 r sg2: 2 l, 2 r sg3: 1 l, 1 rProximal base of neuropodial acicula; proximal neuropodial chaeFCSparse
Inter-acicularinteracicular_muscleMUSac-isg1: 1 l, 1 r sg2: 1 l, 1 r sg3: 1 l, 1 rProximal base of neuropodial acicula; proximal base of notopodial aciculaNone
Anterior oblique muscleAnterior ventral oblique musclesParapodial retractor muscleMUSobA-resg1: 6 l, 5 r sg2: 7 l, 7 r sg3: 7 l, 7 rNeuropodial EC chaeFC and chaetal sac ECs; midline cellsMNspider-ant, MNantacic, MNhose, MNob-contra
Ventral parapodial muscle arcMUSobA-arcsg1: 2 l, 2 r sg2: 4 l, 4 r sg3: 3 l, 3 rBasal lamina next to nerve chord; neuropodial ECs and chaeFC-ECsMNspider-ant, MNantacic, MNob-contra
Medial oblique to mid-parapodiumMUSobA-mppsg1: 1 l, 1 r sg2: 2 l, 2 r sg3: 3 l, 3 rMedial basal lamina next to medial nerve cord and putative radial glia; mid-parapodial ECsMNob-contra, MNhose, MNcrab, others
Mediolateral oblique to mid-parapodiumMUSobA-mlppsg1: 0 l, 0 r sg2: 2 l, 2 r sg3: 0 l, 0 rMedial basal lamina next to mediolateral nerve cord and putative radial glia; mid-parapodial ECsMNhose
Lateral oblique to mid-parapodiumMUSobA-lppsg1: 1 l, 1 r sg2: 2 l, 2 r sg3: 2 l, 2 rLateral basal lamina next to the nerve cord and putative radial glia; mid-parapodial ECsMNob-contra, MNhose, MNsmile, other
Oblique to start of transverseMUSobA-transsg1: 1 l, 1 r sg2: 2 l, 2 r sg3: 2 l, 2 rBasal lamina next to the axochord; ventrolateral ECsMNhose, MNcross
Posterior oblique musclePosterior dorsal oblique muscleNotopodial dorsal oblique muscleMUSobP-notDsg1: 4 l, 4 r sg2: 5 l, 5 r sg3: 5 l, 5 rDorsal longitudinal muscles and ECs near them; ECs of the notopodiumMNpostacic, MNspider-ant
Neuropodial dorsal oblique longMUSobP-neuDlongsg1: 1 l, 1 r sg2: 2 l, 2 r sg3: 2 l, 1 rMUSlong_D; distal neuropodial acFC, neuropodial ECs, and mid-parapodial ECsMNspider-post; MNpostacic
Neuropodial dorsal oblique proximalMUSobP-neuDproxsg1: 3 l, 3 r sg2: 3 l, 3 r sg3: 3 l, 3 rMUSlong_D; dorsal paratroch and ECs around itMNspider-post
Neuropodial dorsal oblique distalMUSobP-neuDdistsg1: 3 l, 3 r sg2: 3 l, 3 r sg3: 3 l, 3 rECs near dorsal paratroch; distal neuropodial acFC, neuropodial ECs and mid-parapodial ECsMNbow
Posterior ventral oblique musclePosterior ventral neuropodial muscleMUSobP-neuVsg1: 5 l, 5 r sg2: 7 l, 7 r sg3: 7 l, 7 rBasal lamina next to VNC; ECs from the distal part of neuropodial acicula to the ventral paratroch areaMNring, MNspider-post, MNob-ipsi, MN_oblique, MNob, many fragments
Posterior ventral notopodial muscleMUSobP-notVsg1: 3 l, 3 r sg2: 3 l, 3 r sg3: 3 l, 3 rBasal lamina next to VNC; septal ECs; EC and chaeFC near distal part of notopodial aciculaMNhose, MNspider-post, MNob-ipsi, MN_oblique, MNladder, MNob-contra, fragments
Oblique to distal inter-acicularMUSobP-Msg1: 0 l, 0 r; sg2: 1 l, 1 r sg3: 1 l, 1 rEach other and basal lamina of the VNC; distal inter-acicular muscleNone
Oblique to body wall near distal inter-acicular and neuropodial YMUSobP-notysg1: 1 l, 1 r sg2: 1 l, 1 r sg3: 1 l, 1 rBasal lamina next to the axochord; septal ECsMNhose, MNpostacic, MNob-ipsi, MNspider-post, fragments
sg0 ventral posterior oblique muscleMUSobPsg0: 2 l, 2 r sg1: 0 l, 0 r sg2: 0 l, 0 r sg3: 0 l, 0 rBasal lamina near VNC and ventral ECs; dorsal ECsMNsmile, MNcrab, MNob-contra, MNmouth
Posterior median oblique muscleDistal inter-acicular muscleMUSobP-isg1: 1 l, 1 r sg2: 1 l, 1 r sg3: 1 l, 1 rNeuropodial EC, chaeFC-EC, acFC-EC; notopodial EC, chaeFC EC; MUSob-postMSparse
Oblique muscle otherMUSobsg0: 1 l, 1 r sg1: 0 l, 0 r sg2: 0 l, 0 r sg3: 0 l, 0 rECs near mouth/lower lip; MUStransMNsmile
Chaetal sac muscleNotopodialNotochaetal next to dorsal obliqueMUSchae-notDobsg1: 3 l, 3 r sg2: 3 l, 3 r sg3: 3 l, 3 rDorsolateral ECs; notopodial chaeFCMNpostatic, various
Dorsal notopodial chaetal sac muscleMUSchae-notDsg1: 1 l, 1 r sg2: 1 l, 1 r sg3: 1 l, 1 rProximal notopodial acicula; notopodial chaeFC (semicircle around dorsal side of notopodial chaetal sac)Sparse
Next to dorsal notopodial chaetal sac muscleMUSchae-notDnsg1: 1 l, 1 r sg2: 1 l, 1 r sg3: 1 l, 1 rDorsolateral ECs; ECs of the notopodial chaetal sac and notopodial chaeFC-ECs; notopodial acFC cellsMNpostacic
Anterior notopodial chaetal sac muscleMUSchae-notAsg1: 1 l, 1 r sg2: 1 l, 1 r sg3: 1 l, 1 rMid-parapodial chaetal sac ECs and MUSobA-che; notopodial ECs, chaeFC-EC and acFCMNspider-ant, MNantacic
Notopodial chaetal muscle under aciculaMUSchae-notAacsg1: 4 l, 4 r sg2: 4 l, 4 r sg3: 4 l, 4 rDistal acFC; all types of chaetal sac cellsMNcrab, MNbiramous, MNwave, MNantacic, MNacic, MNspider-ant, MNbow, fragments
Notopodial retractor muscleMUSchae-notAresg1: 0 l, 0 r sg2: 1 l, 1 r sg3: 1 l, 1 rMid-parapodial ECs; distal notopodial acFC and EC chaeFCSparse
NeuropodialNeurochaetal next to ventral obliqueMUSchae-neuVobsg1: 2 l, 2 r sg2: 3 l, 3 r sg3: 2 l, 2 rProximal neuropodial chaeFC-hemi; ventrolateral ECsMNchae
Neuropodial chaetal muscle under aciculaMUSchae-neuDacsg1: 4 l, 4 r sg2: 5 l, 5 r sg3: 5 l, 5 rMid and distal neuropodial circumacicluar cells; proximal and distal chaeFC cells and chaetal sac ECsMNbiramous, MNchae, MNspider-ant, fragments
Anterior ventral neurochaetal muscle obMUSchae-neuAVosg1: 2 l, 2 r sg2: 3 l, 3 r sg3: 3 l, 3 rNeuropodial mid acFC and proximal chaeFC; distal chaeFC and neuropodial ECsMNarm, MNchae
Anterior ventral neurochaetal muscle transMUSchae-neuAVtsg1: 1 l, 1 r sg2: 1 l, 1 r sg3: 1 l, 1 rDistal neuropodial acFC; ventrolateral ECs next to VLMMNacicX, MNspider-post
Chaetal sac under parapodial retractorMUSchae-Aresg1: 1 l, 1 r sg2: 1 l, 1 r sg3: 1 l, 1 rMid and distal neuropodial chaeFC cellsMNbiramous, MNspider-ant
Transverse muscleTransverse muscleMUStranssg0: 4 l, 5 r sg1: 5 l, 5 r sg2: 9 l, 9 r sg3: 8 l, 9 r pyg: 4 l, 4 rLateral ECs; metatroch; akrotrochMNring, MNhose, MNsmile, MNob-contra, MNladder
Longitudinal muscleDorsolateralDorsolateral muscleMUSlongDsg0: 11 l, 15 r sg1: 13 l, 15 r sg2: 10 l, 10 r sg3: 7 l, 7 rEpidermal cells; MUSobP-neuD, MUSobP-notDprox, MUSobP-notDlong; paratroch; nuchal organMN2, MNring, MNcrab
VentrolateralVentrolateral muscleMUSlongVhead: 9 l, 8 r, 1 m sg0: 7 l, 4 r sg1: 8 l, 12 r sg2: 11 l, 8 r sg3: 7 l, 7 r pyg: 1 l, 1 rBasal lamina lateral of VNC; ventral ECsMN1, MNring, MNcrab, MNsmile, MNspider-ant
AxochordAxochordMUSaxsg1: 4 l, 4 r sg2: 1 l, 1 r sg3: 1 l, 1 rBasal lamina next to VNC; radial glia-like midline; ECsMNax; MNsmile
Digestive system muscleCircular pygidial muscleMUSring-pygpyg1Self; dorsal pygidial ECs; ventral pygidial ECsNone
Pharyngeal muscleMUSph28 l, 30 rBasal laminaNone
Lower lip muscleMUSll5 l, 5 rMetatroch and ECs next to metatrochMNsmile
Head muscleHead appendageAntenna muscleMUSant4 l, 4 rECsSparse
Lyrate muscleMUSly3 l, 3 rBasal laminaNone
Palp muscleMUSpl2 l, 2 rECs and prototroch cover cellsNone
Cirrus muscleMUSci2 l, 2 rECs and basal laminaNone
Other headCheek muscleMUSch3 l, 3 rBasal laminaMNsmile
Plexus muscleMUSpx1ECs and putative gliaSNs with very weak connections
Ventral transverse muscle of prostomiumMUSpr-Vt1Dorsolateral head ECsSparse
Triangle muscleMUStri1 l, 1 rBasal laminaNone
Smooth head musclesMUSmed-head1 l, 1 rHead ECsNone

Classification, left–right stereotypy, and segmental distribution of muscle cell types

The eight main muscle groups could be further subdivided into 53 distinct types (Figure 2, Table 1, Video 1). These 53 types represent anatomically distinct groups of cells that we classified based on their position, shape, motoneuron inputs, and desmosomal partners. All types are composed of segmentally repeated sets with the same number of cells on the left and right body sides, with few exceptions (Figure 2I; Table 1). The precise and repeated features of the anatomy and the type-specific patterns of desmosomal connectivity and motoneuron innervation suggest that these 53 types represent cell types specified by unique development programmes and muscle identity genes, as is the case for example in Drosophila larvae (de Joussineau et al., 2012).

Video 1
All muscles in the 3-day-old Platynereis larva.

Individual muscle cell types are shown by cell type and muscle category. The 12 aciculae are shown in grey for reference. The cell nuclei are labelled by a sphere. The yolk and body outlines are shown in grey. This video can be reproduced by loading the Jasek_et_al.Rproj R project in RStudio and running the code/Video1.R script (Jasek, 2022).

To name each muscle cell type we adapted available names from the annelid anatomical literature (Allentoft-Larsen et al., 2021; Bergter et al., 2008; Filippova et al., 2010; Mettam, 1967). However, our study is the highest resolution whole-body analysis to date, therefore we also had to extend and modify the available nomenclature (Figure 2 and Table 1).

The majority of muscle types (34 out of 53 types corresponding to 502 out of 853 cells) are in the parapodial appendages and form the parapodial muscle complex (Video 2). The parapodial complex of the first segment in the 3-day-old larva has the same major muscle groups as other segments, but fewer muscle cells and some minor subgroups are missing (MUSac-neuDy, MUSac-neuDx, and MUSobP-M). The parapodia of this segment will transform into tentacular cirri during cephalic metamorphosis (Fischer et al., 2010), but at the 3-day-old stage these are bona fide locomotor appendages. The 3-day-old larva also has a cryptic segment (segment 0) (Steinmetz et al., 2011), which lacks parapodia and has few muscle cells (Figure 2I, Table 1).

Video 2
Muscle groups in the parapodial complex.

Reconstruction of all cell groups in the left parapodial complex in the second segment. The neuropodial and notopodial aciculae are shown in black, chaetae are in yellow. The dark-brown cell with the large nucleus is the spinning gland. The yolk outline is shown in grey. Subsequent rotations highlight selected muscle cell types involved in some of the movements in Figure 6—figure supplement 4. CATMAID view of a similar 3D rendering: https://catmaid.jekelylab.ex.ac.uk/11/links/ooeymw3. This video can be reproduced by loading the Jasek_et_al.Rproj R project in RStudio and running the code/Video2.R script (Jasek, 2022).

Reconstruction of the whole-body desmosomal connectome

To analyse how muscles connect to each other and to the rest of the body, we focused on desmosomes. We use a morphological definition of desmosomes as electron-dense structures linking two adjacent cells (desmosomes) or cells to the basal lamina (hemidesmosomes). In non-muscle cells, desmosomes are often anchored by cytoplasmic tonofibrils, thick bundles of intermediate filaments (Figure 1—figure supplement 1, Figure 1—figure supplement 2). In the volume, we sampled and annotated desmosomes and hemidesmosomes and defined the partner cells (or basal lamina fragments) that were interconnected by them. We used this connectivity information and the cell annotations to derive a ‘desmosomal connectome’ containing muscle and other cells and basal lamina fragments (short, locally-traced skeletons) as nodes (vertices) and desmosomes and hemidesmosomes as links (edges). The desmosomal connectome forms a single large interconnected network (Figure 1, Figure 3) after the exclusion of a few small isolated subgraphs (see Methods).

Figure 3 with 1 supplement see all
The desmosome connectome of the 3-day-old Platynereis larva.

(A) The desmosomal connectome coloured by Leiden modules. Nodes represent cells or basal lamina fragments and edges represent desmosomal connections. Node sizes are proportional to weighted degree (sum of all weighted connections). The graph layout was computed by a force-field-based method. Around the network graph, morphological renderings of the cells are shown for each module. Spheres show positions and sizes of nuclei. Grey meshes show the outline of the yolk. The 12 aciculae are shown for segmental reference. Numbers in square brackets after the module names show module id, ordered by module size. (B) Number of nodes, edges, and cells per main cell classes in the desmosomal connectome. (C) Number of cells in each module (ordered by module size) coloured by cell class. (D) Histogram of node degrees (number of connected nodes) for the desmosomal connectome. This figure can be reproduced by running the code/Figure3.R script (Jasek, 2022).

Figure 3—source data 1

The desmosomal connectome graph in igraph format.

https://cdn.elifesciences.org/articles/71231/elife-71231-fig3-data1-v2.zip
Figure 3—source data 2

The desmosomal connectome graph in visNetwork format.

https://cdn.elifesciences.org/articles/71231/elife-71231-fig3-data2-v2.txt
Figure 3—source data 3

The desmosomal connectome graph in html format.

https://cdn.elifesciences.org/articles/71231/elife-71231-fig3-data3-v2.zip

Desmosomes are one of the strongest types of adhesive junction, characteristic of tissues that experience mechanical stress. The position and density of desmosomes in muscle cells in the Platynereis larva suggest that these are the primary junctions mediating force transmission. Other types of junctions including various types of adherens-like junctions (Tilic and Bartolomaeus, 2016) (primarily found in epidermal and chaetal follicle cells) were not considered here. Some adjacent muscle cells are also connected by an unclassified type of adhaerens-like junction (some examples are tagged with ‘junction type 3’ in the volume). These junctions are not in every muscle, are less electron-dense and were not annotated systematically.

We first annotated (hemi-)desmosomes in every 50 layers of the 4847 layer EM dataset. We then manually surveyed each muscle cell and identified their desmosomal partners that were not connected in the first survey. Desmosomes can span multiple 40 nm EM layers (up to 15 layers) and are enriched at the ends of muscle cells indicating that they transmit force upon muscle cell contraction (Figure 1E, Figure 1—figure supplement 1). The (hemi-)desmosomal partners of muscle cells include the basal lamina (34.4% of desmosomes) and a diversity of cell types. These are other muscle cells, glia, multiciliated cells of the ciliary bands (except the prototroch), epidermal cells, and various follicle cells encircling the chaetae (chaetal follicle cells) and the aciculae (acicular follicle cells)(Figure 1—figure supplement 2A, Figure 1—figure supplement 3). Only 2.4% of (hemi-)desmosomes are between muscle cells and 60.5% of these are between different muscle types. This suggests that in the Platynereis larva, desmosomes do not mediate lateral force transmission between muscles of the same bundle. In many desmosomal partner cell types – but not in the muscle cells themselves – we could identify tonofibrils (Figure 1—figure supplement 1 and Figure 1—figure supplement 2).

The full desmosomal connectome is an undirected graph of 2807 interconnected nodes (2095 with a soma) connected by 6961 edges. Six hundred and thirty nodes are fragments of the basal lamina of similar skeleton sizes (see Methods).

Local connectivity and modular structure of the desmosomal connectome

To characterise the structure of the desmosomal network, we first analysed its modularity by the Leiden algorithm, which partitions graphs into well-connected communities (Traag et al., 2019). We detected several dense clusters of nodes or communities in the desmosomal connectome. These communities correspond to anatomical territories in the larval body (Figure 3, Video 3). There are four modules – a left and right ventrolateral and a left and right dorsolateral – consisting of longitudinal muscles spanning all body segments, and associated ciliary band and epidermal cells. Various head and segment-0 muscles constitute another module (Figure 3, Video 3). Two modules contain left–right groups of parapodial and oblique muscles in the second and third segments, connected at the midline through the basal lamina and the median ventral longitudinal muscle (axochord or MUSax) (Lauri et al., 2014; Purschke and Müller, 2006). The other modules include parapodial muscles and chaetal sac cells, in the segmental parapodia. Each module contains muscles and epidermal cells, and various other cell types (Figure 3C).

Video 3
3D visualisation of all cells in the desmosomal connectome coloured by module.

The colour scheme is the same as in Figure 3. The cell nuclei are labelled by a sphere. Individual Z planes from the stack are also shown to indicate the orientation of the serial EM stack. The yolk and body outlines are shown in grey. CATMAID view: https://catmaid.jekelylab.ex.ac.uk/11/links/7ftc5sa. This video can be reproduced by loading the Jasek_et_al.Rproj R project in RStudio and running the code/Video3.R script (Jasek, 2022).

We also generated a grouped connectivity graph where cells of the same type were collapsed into one node (Figure 3—figure supplement 1).

To better understand the organisation of the desmosomal network, we compared it to simulated networks generated by three different stochastic algorithms and a reduced synaptic (neuronal) connectome graph from the same Platynereis larva (Verasztó et al., 2020). We used 1000 simulated graphs for each method and 1000 subsampled graphs from the connectome graphs.

While all graph types have a similar degree and weighted degree distribution (Figure 4A, B), the desmosomal graph stands out as a highly modular and less compact graph with strong local connectivity. This conclusion is supported by several graph measures.

Network statistics of the desmosomal connectome relative to random graphs.

(A) Degree distribution and (B) edge-weight distribution of the desmosomal connectome compared to the synaptic (neuronal) connectome (neuro), scale-free (sf), Erdős-Rényi (erdos), and preferential-attachment (pa) graphs (1000 graphs each). (C) Modularity scores of 1000 weighted scale-free (sf), Erdős-Rényi, and preferential-attachment (pa) graphs relative to the weighted desmosomal and synaptic connectome graphs (1000 subsamples each). (D) Mean diameter scores of 1000 scale-free, Erdős-Rényi, and preferential-attachment graphs relative to the desmosomal and synaptic connectome graphs (1000 subsamples). (E) Mean distance and (F) transitivity (clustering coefficient) scores of 1000 weighted scale-free, Erdős-Rényi, and preferential-attachment graphs relative to the weighted desmosomal and synaptic connectome graphs (1000 subsamples each). (G) Number of 3-member cliques (triangles) and (H) assortativity coefficient. X axes are in sqrt scale for A, B and log10 scale for F, G. Abbreviations: desmo, desmosomal connectome; neuro, synaptic connectome; sf, scale-free; erdos, Erdős-Rényi; pa, preferential-attachment. This figure can be reproduced by running the code/Figure4.R script (Jasek, 2022).

Figure 4—source data 1

Network statistics for the 1,000 simulated scale-free, Erdős-Rényi, preferential-attachment graphs and the 1000 subsampled desmosomal connectome and synaptic connectome graphs.

https://cdn.elifesciences.org/articles/71231/elife-71231-fig4-data1-v2.txt

The modularity scores of the subsampled desmosomal graphs are higher than for any other graph types, including the synaptic connectome (Figure 4C). The desmosomal subgraphs have the largest graph diameter (the length of the longest graph geodesic between any two vertices) and the mean of the distances between vertices (Figure 4D, E), both measures of graph compactness (Doyle and Graver, 1977). The desmosomal and synaptic connectome graphs have the highest transitivity values (clustering coefficient), which measures the probability that nodes connected to the same node are also directly connected to each other (Figure 4F). A similar measure, the number of 3-cliques (three fully connected nodes) also ranks the synaptic graph as first followed by the desmosomal graph (Figure 4G). The assortativity coefficient (between −1 and 1), a measure of the extent to which nodes of similar degree are connected (1 for the preferential-attachment graphs) is lowest for the desmosomal graph, indicating more connections between nodes of different degree (Figure 4H).

Overall, the desmosomal graph stands out as a highly modular graph with a large diameter, large average distance, and high level of local connectivity (cliques). These properties set the desmosomal graph apart from different types of random graphs and the synaptic connectome. We attribute these characteristics to the special organisation of the desmosomal connectome, with all cells only connecting to cells in their immediate neighbourhood forming local cliques, and without long-range connections (e.g. between the left and right body sides).

The tight local connectivity of the desmosomal connectome is also apparent on force-field-based layouts of the graph. In force-field-based layouts, an attraction force is applied between connected nodes, together with a node-to-node repulsion and a general gravity force. As a result, more strongly connected nodes tend to be placed closer to each other.

To analyse how closely the force-field-based layout of the desmosomal connectome reflects anatomy, we coloured the nodes in the graph based on body regions (Figure 5). In the force-field layout, nodes are segregated by body side and body segment. Exceptions include the dorsolateral longitudinal muscles (MUSlongD) in segment-0. These cells connect to dorsal epidermal cells that also form desmosomes with segment-1 and segment-2 MUSlongD cells. These connections pull the MUSlongD_sg0 cells down to segment-2 in the force-field layout (Figure 5D).

Local connectivity of the desmosomal connectome.

(A) Morphological rendering of cells of the desmosomal connectome in the midline (black) and on the right (cyan) and left (orange) side of the body. (B) The same cells in the force-field-based layout of the graph, coloured by the same colour scheme. (C) Morphological rendering of cells of the desmosomal connectome coloured by body region (head, segments 0–3 and pygidium). (D) The same cells in the force-field-based layout of the graph, coloured by the same colour scheme. (E) Morphological rendering of cells of the desmosomal connectome in the neuropodia (orange) and notopodia (cyan). (F) The same cells in the force-field-based layout of the graph, coloured by the same colour scheme. (G) Morphological rendering of muscle cells (red), aciculae (black), and acicular follicle (acFC) cells (cyan). (H) The same cells in the force-field-based layout of the graph, coloured by the same colour scheme. Spheres in the morphological rendering represent the position and size of cell nuclei. Nodes in the connectivity graphs represent cells with their sizes proportional to weighted degree. This figure can be reproduced by running the code/Figure5.R script (Jasek, 2022). An interactive html version of the graph with node labels is in the GitHub repository supplements/Fig5_desmo_connectome_seg_interactive.html.

Nodes corresponding to the neuropodial (ventral) and notopodial (dorsal) parts of the parapodia also occupy distinct domains (Figure 5E, F). The 12 aciculae and the acicular follicle cells also occupy positions in the force-field map paralleling their anatomical positions (Figure 5G, H). The observation that node positions in the force-field layout of the graph recapitulate the anatomical positions of the nodes reflects the local connectedness (neighbours only) of the desmosomal connectome. This is different from the synaptic connectome where long-range neurite projections can link neurons at two different ends of the body (Verasztó et al., 2020).

High muscle diversity and strong connectivity of aciculae in the parapodial complex

We further queried the desmosomal connectome to better understand the organisation and movement of the annelid larval body. We focused on the parapodia as these contain the largest diversity of muscle types and associated cells (Video 2).

In the segmented larva, each parapodium has a dorsal and a ventral lobe (notopodium and neuropodium). Each lobe is supported by an internal chitinous rod called acicula (Hausen, 2005) that is ensheathed by acicular follicle cells (Figure 1—figure supplement 3A, C). Each parapodial lobe also contains a chaetal sac consisting of bristle-producing chaetoblasts, chaetal follicle cells, epidermal cells (Figure 1—figure supplement 3B, C), and bristle mechanosensory neurons (chaeMech) (Verasztó et al., 2020). The chaetal follicle cells ensheathing the chaetae have four types arranged in a proximo-distal pattern (Figure 6—figure supplement 1). The aciculoblasts, chaetoblasts, and the various follicle cells serve as anchor points for several major parapodial muscle groups (Figure 6E–I, Figure 1E, Video 4).

Figure 6 with 4 supplements see all
Cell-type diversity and connectivity of the parapodial muscle complex.

(A) Morphological rendering of all cells of the desmosomal connectome with a weighted degree >10. (B) Same cells coloured with a colour scale proportional to the cell’s weighted degree, ventral and (C) lateral views. (D) The desmosomal connectome with node-colour intensity and node size proportional to node weighted degree. (E) Weighted degree of the most highly connected cells in the desmosomal connectome, arranged by cell class. Colour scale in E also applies to B–D. (F) Morphological rendering of the outlines of muscle cell types that directly connect through desmosomes to the acicular follicle cells. Anterior view of a transverse section showing neuro- and notopodia in the left side of the second segment. (G) Desmosomal connectivity graph of acicular follicle cells and their partners. Nodes represent groups of cells of the same cell type. Aciculae and acicular follicle cells are separated into neuropodial and notopodial groups. (H) Morphological rendering of the outlines of muscle cell types that directly connect through desmosomes to the chaetal follicle cells. Anterior view of a transverse section showing neuro- and notopodia in the left side of the second segment. (I) Desmosomal connectivity graph of chaetal follicle cells and their partners. Nodes represent groups of cells of the same cell type. Chaetae and chaetal follicle cells are separated into neuropodial and notopodial groups. Edge thickness is proportional to the number of desmosomes connecting two cell groups. This figure can be reproduced by the code/Figure6.R script (Jasek, 2022).

Figure 6—source data 1

Source data for panel E of Figure 6.

Text file listing weighted degree values for the most highly connected cells in the desmosomal connectome, arranged by cell class.

https://cdn.elifesciences.org/articles/71231/elife-71231-fig6-data1-v2.txt
Video 4
3D visualisation of selected muscle groups and their desmosomal connections.

Morphological renderings of selected muscle groups and partner cells and the desmosomes (in red) that connect them. The video illustrates that desmosomes often occur at the most distal ends of muscle cells. This video can be reproduced by loading the Jasek_et_al.Rproj R project in RStudio and running the code/Video4.R script (Jasek, 2022).

The aciculae and associated structures in the parapodial complex represent highly connected components in the desmosomal connectome (Figure 6E–G). The cells with the largest mean degree and weighted degree belong to the parapodial complex. When we plot all cells in the desmosomal connectome with a colour-transparency inversely proportional to node weight, the parapodial muscle complex is highlighted (Figure 6B, C).

We identified 16 muscle types with connections to the aciculae or to acicular follicle cells (Figure 6F, G; Figure 6—figure supplement 2; Figure 6—figure supplement 3). These include acicular muscles (MUSac), some chaetal muscles (MUSchae) that link the aciculae to the chaetae and oblique muscles (MUSob) that link the aciculae to the dorsal longitudinal muscles (MUSlongD).

Several acicular muscles attach on one end to the proximal base of the aciculae and on the other end to the paratrochs and epidermal cells. Oblique muscles attach to the basal lamina, epidermal, and midline cells at their proximal end, run along the anterior edge of parapodia and attach to epidermal and chaetal follicle cells at their distal tips (Figure 6—figure supplement 2, Videos 2 and 4). Both of these muscle groups are involved in moving the entire parapodium. Acicular muscles move the proximal tips of the aciculae, while oblique muscles move the parapodium by moving the tissue around the chaetae and the aciculae. All acicular movements also correspond to parapodial movements. Chaetae are embedded in the parapodium and therefore move with it, but the chaetal sac muscles can also independently retract the chaetae into the parapodium or protract them and make them fan out.

The notopodial and neuropodical aciculae have distinct muscle partners indicating that these structures could move independent of each other. Two muscles (MUSac-i and MUSac-neuDx) link the noto- and neuropodial aciculae within the same parapodium to each other, suggesting force coupling between the two aciculae (Figure 6E, G). We analysed each muscle group for their desmosomal partners and spatial orientation and inferred the possible movements of the aciculae upon their contraction. This revealed several possible acicular movements which we termed extension, flexion, pivoting, abduction, chaetal retraction, and jostling (Figure 6—figure supplement 4, Video 2, Video 4). In addition, some muscles connect to epidermal cells in the parapodium and through moving these epidermal cells can indirectly also move the aciculae (i.e. ‘pulled by the skin’; e.g. MUSobP-neuV) (Video 4).

The movement of the parapodia during crawling also requires connections to the axochord muscle (MUSax) at the midline. Animals with an ablated axochord show impared crawling indicating a structural role for this muscle (Lauri et al., 2014). A high density of hemidesmosomes between the axochord and the extracellular matrix dorsal to the neuronal midline agrees with this (Figure 1—figure supplement 1B, Figure 1—figure supplement 2B, D). The extracellular matrix on the midline also serves as an attachment site for the proximal ends of anterior and posterior oblique muscles and the radial-glia-like midline cells that are rich in tonofibrils.

Acicular movements and the unit muscle contractions that drive them

The desmosomal connectome suggests that each acicula can have complex and extensive movements and the two aciculae can move independently of each other, with some coupling (e.g. through the inter-acicular muscles). To observe acicular movements in live animals, we first imaged 4-day-old crawling Platynereis larvae with differential interference contrast (DIC) optics. The organisation of the musculature in 3- and 4-day-old three-segmented larvae is very similar (Figure 7—figure supplement 1), therefore we could relate movements in the more active 4-day-old larvae to the EM data.

We used the toolbox DeepLabCut (Mathis et al., 2018) to train a deep residual neural network using sample video frames to learn, track, and label 26 body parts of the larvae. This allowed us to track the 12 individual aciculae and their relative angles, both to the body midline (Figure 7A) and to one another (Figure 7B). We found that the notopodial and neuropodial aciculae within one parapodium exhibit differences in movement velocity and angles relative to one another resulting in a range of angles and positions over time (Figure 7, Video 5). During the crawl cycles, the aciculae first draw inwards and forwards, diagonally towards the head, then they tilt in a posterior direction, the proximal tip of the neuropodial acicula travelling slightly faster to open up the pair, creating the larger (40–50°) inter-acicular angle (Figure 7E–G). This is followed by a ‘piston’ movement whereby the now parallel aciculae are pushed outwards, then the proximal tips tilt rostrally again, to move the chaetae back against the trunk, propelling the animal forward on that side and completing that acicular cycle.

Figure 7 with 1 supplement see all
Acicula tracking, muscle contractions, and desmosomal connectivity.

(A) Acicular position over time in all six parapodia analysed from video tracking data to indicate gait of the larva over time. Angles between the anterior–posterior bodyline connecting the mouth and the hindgut (dotted line, panel C) and each notopodial acicula were normalised (scale 0–100) within each parapodium’s angular range of motion. White areas of the heatmap represent aciculae held fully back, close to the trunk. Black areas represent aciculae extended to their extreme forward position. Arrows indicate time points at which panels C–G were taken. (B) Inter-acicular angle between notopodial and neuropodial aciculae over time. Frames of the video sequence are labelled to show (C) the larva relaxed, (D) showing a partial startle response, (E, F) during crawl cycle 2 where opposite acicula pairs diverge, extend rostrally and inward towards the trunk, causing the larva to bend left or right. (G) From crawl cycle 1, acicular positions from one example acicular cycle (lasting 720 ms) are plotted for the left parapodium in segment 2, as the larva crawls forward. Points mark the proximal tips of the two aciculae, connected by coloured lines to indicate time. Finer light/dark grey straight lines show the relative positions and angles of the distally projecting aciculae. Below, the inter-acicular angles are plotted for the same parapodium over this time period. (H) Live imaging of MUSac-neuPV, (I) MUSobP-neuV, and (J) MUSobA-re contraction and neuropodia displacement. In H, I, and J, the left panel shows the GCaMP6s signal, the right panel shows the differential interference contrast (DIC) channel, the top and bottom panels show two frames from a video (t1, t2). (K) Desmosomal connectivity of MUSac-neuPV. Skeletons of MUSac-neuPV and their desmosomal partners in segment-2. (L) Desmosomal connectivity of MUSobP-neuV. Skeletons of MUSobP-neuV and their desmosomal partners in segment-2. (M) Desmosomal connectivity of MUSobA-re. Skeletons of MUSobA-re and their desmosomal partners in segment-2. Abbreviations: ac-not, notopodial acicula; ac-neu, neuropodial acicula; BL, basal lamina; EC, epidermal cell; sg, segment.

Video 5
Acicular movements in crawling and startling Platynereis larvae.

Differential interference contrast (DIC) video of a crawling larva and a startling larva (end of the video) with the position of the aciculae tracked by DeepLabCut.

This analysis demonstrates that the acicular pairs can have different relative positions with inter-acicular angles changing between −25° and 50° in one parapodial cycle (Figure 7B, G). These relative changes occur in each segment during crawling, propagating from posterior to anterior through an undulatory gait cycle (Figure 7A).

In order to directly visualise muscle contractions, we imaged calcium transients in larvae ubiquitously expressing GCaMP6s. Larvae held between a slide and coverslip display spontaneous contractions of different muscle groups (‘twitches’) revealed by increased GCaMP fluorescence. Simultaneous imaging in the DIC channel allowed us to visualise the movement of the aciculae and parapodia (Figure 7H–J; Video 6).

Video 6
Calcium imaging of muscle contractions in Platynereis larvae.

Spontaneous contractions of individual muscle groups as visualised by the confocal imaging of calcium signals reported by GCaMP6s fluorescence. The larvae were also imaged in the differential interference contrast (DIC) channel to reveal the movements of the parapodia.

The contraction of the posterior ventral neuropodial acicular muscle (MUSac-neuPV) pulls the proximal end of the acicula caudally, inducing an abduction (Figure 7H; Figure 6—figure supplement 4E). These muscles connect to the proximal end of the acicula via acicular follicle cells and anchor to the basal lamina, epidermal cells and ciliated cells (paratroch) at their other end (Figure 7K).

The contraction of the posterior ventral neuropodial muscles (MUSobP-neuV) induces an inward movement of the distal tip of the acicula (‘flexion’, Figure 6—figure supplement 4B) leading to the alignment of the parapodia and chaetae with the longitudinal body axis (Figure 7I). These muscles connect to the distal end of the aciculae via acicular follicle cells and also connect to epidermal cells in the neuropodium. At their other end, MUSobP-neuV cells are anchored to the basal lamina (Figure 7L).

The MUSac-neuPV cells can contract simultaneously with the parapodial retractor muscles (MUSobA-re), inducing parapodial retraction and a tilt of the aciculae with their proximal tip moving rostrally (Figure 7J). The retractor muscles form desmosomes on chaeFC-EC (epidermal chaetal follicle) and epidermal cells at their distal end in the neuropodium and anchor to ventral midline cells and the basal lamina at their other end (Figure 7M). The ectodermal ventral midline cells bear similarities to radial glia (Helm et al., 2017) but lack cilia.

We could only observe these spontaneous individual muscle contractions in non-crawling larvae. During a startle response or crawling cycles, many muscles contract rapidly (Bezares-Calderón et al., 2018) and we were not able to spatially and temporally resolve these to individual muscle groups.

Combined analysis of synaptic and desmosomal networks

The availability of full desmosomal and synaptic connectomes (Verasztó et al., 2020) for the same Platynereis larva allowed us to analyse how individual motoneurons could influence muscles and associated tissues. Motoneuron activation is expected to induce postsynaptic muscle contraction, which will exert forces on the desmosomal partners of the muscle. By combining desmosomal and synaptic connectomes we can infer the impact of motoneuron activation on tissue movements (Figure 8).

Figure 8 with 1 supplement see all
Combining synaptic and desmosomal connectomics.

(A) Graph representation of motoneuron synaptic inputs to the desmosomal connectome. The 11 main motoneuron groups are shown. (B) Sankey diagram of the synaptic innervation of muscle types by motoneurons. The width of the bands is proportional to the number of synapses from a motoneuron class to a muscle class. Only connections with >9 synapses are shown. Cell types are coloured by Leiden modules determined for this grouped synaptic graph. (C) MNcrab motoneuron synaptic inputs to muscle cells (orange) in the desmosomal connectome and the desmosomal partners of these muscle cells (green). (D) Morphological rendering of the four MNcrab motoneurons. (E) The postsynaptic muscle partners (shades of red) of the MNcrab neurons (blue). (F) The postsynaptic muscle partners (shades of red) of the MNcrab neurons (blue) and the desmosomal partners of the innervated muscles (shades of green). (G) MNbow motoneuron synaptic inputs to muscle cells (orange) in the desmosomal connectome and the desmosomal partners of these muscle cells (green). (H) The two MNbow motoneurons. (I) The postsynaptic muscle partners (shades of red) of the MNcrab neurons (blue). (J) The postsynaptic muscle partners (shades of red) of the MNcrab neurons (blue) and the desmosomal partners of the innervated muscles (shades of green). In D–F and H–J, the yolk outline is shown in grey for reference. In A, C, and G the desmosomal links are not shown for clarity.

We analysed 11 motoneuron types that are well developed in the 3-day-old larva (Bezares-Calderón et al., 2018; Verasztó et al., 2020). These collectively provide broad innervation across the entire muscle network (Figure 8A). For this analysis, we omitted motoneurons which could not be assigned to a well-annotated cell-type category. To analyse the innervation and desmosomal connectivity of individual motoneuron types, we first plotted the synaptic pathways from the 11 motoneuron groups to the 25 out of 53 muscle groups that receive synaptic innervation (Figure 8B). All motoneurons have multiple muscle targets and most muscles receive synapses from more than one motoneuron.

To characterise the tissue range of each motoneuron type, we first queried their postsynaptic muscle partners in the synaptic connectome and then retrieved all desmosomal partners of these muscles (Figure 8F and J). We derived such combined synaptic-desmosomal graphs for 10 motoneuron classes (Figure 9). Plotting the skeleton reconstructions of these cells highlights the extent of the tissue under the influence of a motoneuron class (Figure 8C–J and Figure 8—figure supplement 1, Video 7). For example, when the most highly connected muscle in the parapodial complex (MUSchae-notAac) contracts, it is expected to transmit the forces of the contraction through desmosomes to up to nine distinct non-muscle cell types (Figure 9).

Combined synaptic–desmosomal graphs of motoneurons.

Synaptic connections (orange arrows) of motoneuron classes to muscles and desmosomal links (blue edges) of the innervated muscle cells. Combined synaptic-desmosomal graph for (A) MNantelope, (B) MNacicX, (C) MNbiramous, (D) MNbow, (E) MNcrab, (F) MNhose, (G) MNladder, (H) MNspider-ant, (I) MNspider-post and (J) vMN1 and vMN2 motoneurons. Only the muscle cells directly innervated by the motoneurons are shown.

Figure 9—source data 1

A zip archive of CATMAID json files.

The graphs can be loaded in the CATMAID graph widget at https://catmaid.jekelylab.ex.ac.uk.

https://cdn.elifesciences.org/articles/71231/elife-71231-fig9-data1-v2.zip
Video 7
3D visualisation of motoneurons, their muscle targets, and the desmosomal partners of the muscles.

The first half of the video shows the combined synaptic and desmosomal connectome of different types of motoneurons. The muscles postsynaptic to the motoneurons are shown in orange. The desmosomal partners of these muscle cells are shown in green. The second half of the video shows all motoneurons from the 12 analysed motoneuron classes, their muscle partners, and the desmosomal partners of those muscles. A similar visualisation is also available in CATMAID where it can be explored interactively: https://catmaid.jekelylab.ex.ac.uk/11/links/8lzcofe. This video can be reproduced by loading the Jasek_et_al.Rproj R project in RStudio and running the code/Video7.R script (Jasek, 2022).

Next we focused on the acicular muscle complex and highlighted each muscle and their desmosomal partners that are under the influence of a motoneuron cell type. There are eight motoneuron cell types that innervate the parapodial complex (MNcrab, MNbiramous, MNwave, MNspider-ant, MNspider-post, MNhose, MNchae, and MNring). Each of these innervates a unique combination of muscle targets – not a single muscle type (Figure 10). This suggests the concerted activity of anatomically distinct muscle types during the parapodial crawl cycle and other appendage movements.

Postsynaptic muscle targets in the parapodial complex of different motoneurons.

Morphological rendering of postsynaptic muscle targets (red-purple) within the parapodial complex of (A) MNcrab, (B) MNbiramous, (C) MNwave, (D) MNspider-ant, (E) MNspider-post, (F) MNhose, (G) MNchae, and (H) MNring motoneurons. The desmosomal partners of the innervated muscles are also shown (green). Only the muscles in the left parapodium of segment 2 and the motoneurons innervating this parapodium are shown.

Discussion

Here, we reconstructed the somatic musculature and attached tissues in the nereid larva and showed how adhesion networks can be analysed for an entire body. We call this approach desmosomal connectomics, where the desmosomal connectome comprises all cells and extracellular structures (basal lamina, aciculae, and chaeta) and their desmosomal connections.

Our reconstructions revealed the high diversity of muscle cell types in the larva. The complexity of the nereid musculature contrasts to the relatively simple muscle architecture in the nematode C. elegans and the tadpole larva of the tunicate Ciona intestinalis. The C. elegans hermaphrodite has 95 body wall muscles all with a similar rhomboid shape (Gieseler et al., 2017). In the C. intestinalis larva, there are 36 muscle cells of 10 types (Nakamura et al., 2012). In contrast, the 3-day-old nereid larva has 853 muscle cells belonging to 53 types.

A large number of muscles engage in the movement of the parapodial muscle complex. This suggests that the diversity of nereid somatic muscles is due to the presence of an endoskeleton. Polychaete annelids are the only animals outside the tetrapods that have trunk appendages rigidified by an endoskeleton. Aciculae probably evolved in the stem group of errant annelids around the Early Ordovician (Parry et al., 2019; Vinther et al., 2008) indicating the deep ancestry of these structures, predating tetrapod limbs. In contrast to tetrapod limb bones, the aciculae are independent skeletal elements without joints. Their function is nevertheless similar in that they provide attachment sites for limb muscles and rigidity in the appendages during movement. In each parapodium, the two aciculae are moved by unique and shared sets of muscles and change their relative angle within a crawling cycle.

Beyond the complexity of the somatic musculature, our global analysis also revealed the diversity of desmosomal partners of muscle cells. These comprise over 10 different cell types, ranging from epidermal and various follicle cells through pigment cells to ciliary band cells. The stereotypy of the desmosomal connections across segments and body sides suggests that – analogous to the synaptic connectome – the desmosomal connectome develops through a precisely specified connectivity code between cell types. Some of these connections have a clear functional relevance, as the attachment of muscles to movable chitin structures. Others are more intriguing, such as the connections to locomotor ciliary cells. These may be due to proximity and the need for muscles to attach to something, alternatively, may mediate hitherto unexplored mechanistic interactions between muscular and ciliary locomotion or represent signalling connections (Green and Gaudry, 2000).

The whole-body desmosomal connectome described here in combination with the synaptic connectome from the same volume (Verasztó et al., 2020) provide the foundation for understanding locomotor control in the nereid larva. In a crawling larva, the contraction of a large number of muscles is recruited in a precise temporal order in each limb to execute a full appendage cycle. The limb movements are organised into a gait sequence, starting from the posterior-most segment (Figure 7). Our imaging of spontaneous muscle twitches (Figure 8), prediction of unit acicular movements (Figure 6—figure supplement 4) and analysis of motoneuron innervation (Figure 9) are the first steps towards understanding parapodial movements and their neuronal control. The availability of a laboratory culture, whole-body connectome and a growing repertoire of genetic tools make the nereid larva a promising experimental system for the in-depth analysis of locomotor control.

We argue that in Platynereis, as well as in other animals, understanding and accurately modelling locomotion will not only require analysing nervous activity and synaptic contacts, but also comprehensive maps of the adhesive networks of muscle systems.

Methods

Cell reconstruction and annotations

We reconstructed all muscle and other cells by skeletonisation in CATMAID (Saalfeld et al., 2009). For each cell, we marked the position of the soma. Muscle cells were identified by the presence of striated myosin filaments. Developing muscles with a morphology similar to differentiated muscle cells but lacking myosin fibres were not annotated with a muscle cell type. Short segments of the basal lamina were also reconstructed as skeletons. Each basal lamina segment has at least two desmosomes and a cable length <63,700 nm. Using such relatively short and often branched fragments allowed us to only focus on basal-lamina-mediated connections for nearby cells. The basal lamina spans the entire body and runs between the ectoderm and the mesoderm. When we consider the basal lamina as one giant skeleton, it is the highest ranking node in the network and distorts the spatial layout (not shown).

We classified muscle and other cells into cell types based on their position, morphology, ultrastructural features, desmosomal connectivity, and synaptic inputs. The complete classification of neuronal cells for the same volume is described in Verasztó et al., 2020.

We annotated all cells by their cell-type categories (Table 1), the position of the soma in a body segment (episphere, segment_0, segment_1, segment_2, segment_3, pygidium), and body side (left_side, right_side), and specific ultrastructural features (e.g. vacuolar ER, villi, dense cored vesicles, etc.). Tonofibrils were tagged as ‘black fibers’ in every 50th layer, or when they were encountered during tracing. Every cell containing at least one ‘black fibers’ tag was also annotated with the ‘black fibers’ annotation.

Desmosomal connectome reconstruction

In order to mark muscle-attachment sites, we added a new connector type to CATMAID: the desmosome connector (besides synapse, gap junction, and other types). Within CATMAID, skeletonslink to each other through so-called connector nodes. Each one acts as a hub so that two skeletons can connect to each other at the same location. The type of such a link is determined by the relation attached to it. For desmosome connectors, two skeletons are allowed to connect to a connector node using the desmosome_with relation. We also updated various APIs and CATMAID front-end tools to support the new desmosome connector type. The connectors can then be displayed and analysed with other CATMAID widgets, including the Connectivity, the 3D View, and the Graph widgets. In Natverse, synaptic and desmosomal connectomes can also be separately analysed.

Desmosomes can be identified as dark electron-dense plaques on cell membranes with a lighter electron-dense extracellular core area (desmoglea). Intracellular intermediate filaments frequently connect to the electron-dense desmosome plaques. Hemidesmosomes were marked as connections between a cell and the basal lamina skeleton. To sample desmosomes across the body, we first sampled every 50th layer in the 4847-layer EM volume and marked every desmosome and hemidesmosome in those layers, on all cells and basal lamina. We then identified all muscle cells with less than two desmosomes from this first sampling and individually identified and marked their desmosomes.

Network analyses

The desmosomal connectome graph was generated with the R code desmo_connectome_graph.R available in the code repository. First, we extracted all annotated desmosomes (12,666 desmosomes) from CATMAID using CATMAID’s ‘connectors’ API endpoint. Next we retrieved all skeletons connected to the desmosome and generated a graph (2903 nodes). The largest connected component was extracted and used as the final desmosomal connectome graph (2807 nodes). Ninety-six individual nodes or nodes forming smaller clusters were removed. The final network (CATMAID annotation: desmosome_connectome) contains 2807 skeletons of which 2095 are cells with a tagged soma. The grouped graph was generated from the desmosomal graph graph by merging nodes of the same cell type into one node.

For force-field-based clustering, we exported the graph in gexf format with rgexf::igraph.to.gexf and imported it into Gephi 0.9. Force-field clustering was carried out with the Force Atlas tool in Gephi (0.9.2) The inertia was set to 0.1, repulsion strength was 35, attraction strength was 10, maximum displacement was 5, gravity was 50, speed was 5, and the ‘attraction distribution’ option was selected. The ‘auto stabilise function’ was off. Towards the end of the clustering the ‘adjust by sizes’ option was also selected. To prevent node overlap, we then run the ‘Noverlap’ function.

The new layout was exported from Gephi with normalised node coordinates as a gexf file and reimported into Rstudio with rgexf::read.gexf. The graph was then visualised with the visNetwork package with the coordinates obtained from the gexf file. Colouring was based on annotations obtained from CATMAID for each skeleton catmaid::catmaid_get_annotations_for_skeletons. The same Gephi layout could also be imported into the CATMAID Graph widget in.graphml format.

To detect modules, we used the Leiden algorithm (Clauset et al., 2004; Traag et al., 2019) (leiden::leiden method in R) with the partition type ‘RBConfigurationVertexPartition’ and a resolution parameter of 0.3.

To compare the desmosomal connectome to stochastic graphs, we used an R script (code/Figure4.R in the repository Jasek, 2022) to generate 1000 each of Erdős-Rényi, scale-free and preferential-attachment graphs with the same number of vertices and edges as the desmosomal graph. To obtain weighted graphs, we assigned the edge weights from the desmosomal graph to these stochastically generated graphs. We also generated 1000 subsamples of the desmosome graph, each with 100 nodes randomly deleted. In addition, we also generated 1000 subsamples of a reduced neuronal connectome graph from the same Platynereis larva (Verasztó et al., 2020).

Motoneuron innervation

Motoneurons and their synaptic connections were reconstructed as described in Verasztó et al., 2020. No gap junctions were observed in our dataset.

Data and code availability

All reconstructions, annotations, and EM images can be queried in a CATMAID project at https://catmaid.jekelylab.ex.ac.uk (project id: 11), together with the synaptic connectome (Verasztó et al., 2020). The data can also be queried by the R package Natverse (Bates et al., 2020). All analysis scripts used are available at https://github.com/JekelyLab/Jasek_et_al (commit b661583, Jasek, 2022; copy archived at swh:1:rev:b661583e56e482f7103629e5cdf23eba12813264).

Anatomical visualisation and preparation of figures

To visualise the morphology of skeletons and soma positions in 3D, we used either the CATMAID 3D view widget or the Natverse package (Bates et al., 2020) in R with the rgl plot engine.

Videos 14 and Video 7 were generated in Rstudio (2022.07.01). Figures were done either in Rstudio (Figures 16, Figure 1—figure supplements 2 and 3, Figure 8—figure supplement 1) or in Adobe Illustrator (various versions) or Inkscape with some panels generated in CATMAID or Rstudio. All R code used is available at https://github.com/JekelyLab/Jasek_et_al, as a single R project (commit b661583).

Live imaging

To image muscle contractions, we used larvae expressing GCaMP6s. Zygotes were microinjected with GCaMP6s mRNA as described (Bezares-Calderón et al., 2018). For imaging, we used a Zeiss LSM 880 confocal microscope with a ×40 C-Apochromat ×40/1.2 W Korr FCS M27 water-immersion objective and a 488 nm laser.

Crawling 4-day-old larvae were imaged with a Leica DMi8 microscope with DIC optics and a V1212 phantom vision research camera.

Phalloidin staining

Three-day-old (72–77 hr after spawning) and 4-day-old (96–100 hr after spawning) larvae were relaxed in isotonic MgCl2 solution (0.34 M, in distilled water) mixed 1:1 with the culture seawater. 4% PFA (paraformaldehyde in 0.1 M phosphate-buffered saline [PBS], pH 7.2, with 0.1% Tween-20) was added 1:1 to fix larvae for 15 min at room temperature (RT) on a rocking board (effective fixation concentration = 2% PFA). Specimens were washed six times for 10 min in 0.1 M PBS with 0.5% Tween-20. Muscles were labelled with 0.33 µM Alexa Fluor 633 phalloidin (Invitrogen) in 0.1 M PBS with 0.5% Triton-X and 0.025% BSA for 1.5 hr at RT on a rocking board. Incubation and all further steps were carried out in the dark. Following six 10 min washes with 0.1 M PBS, 0.5% Tween-20 specimens were mounted in Fluoromount G with DAPI (Invitrogen). At least three specimens per stage were imaged on a Leica TCS SP8 CLSM microscope with a 405-nm diode laser for DAPI and a 633-nm HeNe laser for phalloidin labelling. We used the HyD-hybrid detectors and a ×63 or ×100 oil immersion objective. Image stacks were exported as.lif files and further analysed in Imaris (Bitplane).

Tracking of acicular movements

We used the toolbox DeepLabCut (Mathis et al., 2018) to train a deep residual neural network (ResNet-50). In 30 sample video frames, we manually labelled 26 body parts of the larva, including the end of the pharynx and the developing proctodeum (posterior gut lumen), in addition to two points along the larva’s 12 aciculae; the proximal tip and a more distal point. The network ran 150,000 training iterations before we evaluated it and removed outliers, making corrections for refinement. Afterwards, we merged and retrained the network to improve tracking accuracy. Two high-speed video sequences were then analysed by the trained network using this DeepLabCut toolbox. The first recording incorporated a partial startle response to physical disturbance and three complete crawling cycles. The second recording featured a complete startle response. We then created videos with labelled body parts using the filtered predictions of the network, which appear as coloured dots, some of which (e.g. proximal and distal points of the same acicula) are connected by black ‘skeleton’ lines. A line also connects the pharynx and proctodeum, providing a longitudinal bodyline axis (see also dotted line in Figure 7C). Once satisfied with the accuracy of the tracking, we analysed the xy positions and relative angles of aciculae from the tracking data (Figure 7—source data 1). To reveal patterns of gait as the larva crawled, the angles between each notopodial acicula and the larva’s longitudinal axis were normalised (scale 0–100) within each acicula’s own extremes of movement. This normalisation was necessary due to the different angular placements and ranges of motion between the parapodia along the segments of the body. In addition, we calculated the angles between acicular pairs within the same parapodium to assess the degree to which they remain parallel and at which times they diverge or move independently.

Data availability

All EM, tracing and annotation data are available at https://catmaid.jekelylab.ex.ac.uk. All code is available at https://github.com/JekelyLab/Jasek_et_al (copy archived at swh:1:rev:b661583e56e482f7103629e5cdf23eba12813264).

The following data sets were generated
    1. Jasek S
    2. Verasztó C
    3. Shahidi R
    4. Jékely G
    (2021) CATMAID
    ID HT-4_Naomi. Desmosomal connectome tracing and annotation data.

References

    1. Albertson DG
    2. Thomson JN
    (1976) The pharynx of Caenorhabditis elegans
    Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 275:299–325.
    https://doi.org/10.1098/rstb.1976.0085
    1. White JG
    2. Southgate E
    3. Thomson JN
    4. Brenner S
    (1986) The structure of the nervous system of the nematode Caenorhabditis elegans
    Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 314:1–340.
    https://doi.org/10.1098/rstb.1986.0056

Decision letter

  1. Kristin Tessmar-Raible
    Reviewing Editor; University of Vienna, Austria
  2. Marianne E Bronner
    Senior Editor; California Institute of Technology, United States

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Desmosomal connectomics of all somatic muscles in an annelid larva" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Marianne Bronner as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

All reviewers were impressed by the large dataset and agree that the paper is in principle suitable as a resource article for eLife. However, there had been some discussions on the required extent of data re-analysis, ultimately connected to the question what kind of biological insights might be possible to answer with this dataset as it stands and what is important for biological meaningful insights in the future. Based on this discussion, we concluded on the following requests:

1. Adding information on the presence/absence of gap junctions. This would provide important information on the possible cellular communication.

2. Adding the consideration of the vector-like nature of muscle cells and the likely different functions of desmosomes depending on their relative position in the muscle cell to the manuscript. The suggestion is to filter your data on the level of each muscle cell, which subsequently should allow to tell apart three regions: front and back end, where the fibers exert contraction force, and sides, where desmosomes in-between fibers (or between fibers and epidermis) somehow stabilize the muscle or bodywall. This should provide important detailed information as to how the muscles could exert their forces.

There had been some discussion as to how feasible the addition of the requested information of points 1/2 is within a reasonable timeframe. If this appears to be an unreasonable request, we would ask the authors to get back to us with a suggestion how they could discuss this limitation of their data analysis.

3. improved visualisation of the data, especially details on the acicular movements:

The descriptions of the aciculae are difficult to follow. A simulation/video showing how the respective muscle groups lead to different movements of the acicula in 4D or a detailed annotation of Video 4 would be highly informative. Please also add information on where (relative to the acicular base) the epithelial cell (or point in the basal lamina) is located.

Please provide all annotations to a level that is understandable to the interested non-annelid expert.

4. Reformulations of different parts of the manuscript have been requested by all reviewers. There are two rationales behind this request.

A) Improvement of the understandability. Please see the individual "recommendation for the authors":

Reviewer 2: see "recommendation for the authors"

Reviewer 3: questions mentioned under point 1

B) A clearer vision of what kind of biologically relevant questions can be answered with the dataset in the future.

Reviewer #1 (Recommendations for the authors):

The timeliness, depth and quality of analysis and importance towards a full-body understanding of locomotion makes this manuscript an important and interesting work for scientists of a range of different fields, and I therefore think that is highly suited for publication in eLife. The manuscript generally reads well, although some parts would benefit from more precise formulations, clarification and editing such as restructuring or removing technical details.

– It would be interesting to know if muscle cells are not only physically connected by desmosomes but also electrically by gap junctions. It would be interesting if the authors could add data/information about this point to the manuscript.

– I find the possible acicular movements ('abduction,…'p.8, last paragraph and Figure 6-supp3) very difficult to follow. One problem might be that in Figure 6-supp3, the body axes are mostly missing. In addition, it is inherently difficult to understand spatial and temporal changes using a two-dimensional schematic. If possible, a simulation/video showing how the respective muscle groups lead to different movements of the acicula would be highly informative.

– Related to the previous point, I have difficulties to follow the relative 4-dimensional movements of the aciculae described in Figure 7G, and a 4D animation (if possible) of the relative movements of the aciculae would certainly help the reader. Alternatively, annotating Video 4 might help understanding acicular movements better.

Reviewer #2 (Recommendations for the authors):

1) This paper is rich in data and anatomical details, and would be of benefit to scientists who primarily work on other model systems. An obstacle to this goal is that the current draft is dense, making it a slow read (I learned a lot, after several goes). Reorganization can improve this article's readability for a broader audience:

– Some jargon-like technical details distract readers from focusing on main ideas. I would relegate them to Methods (Example: how authors use CATMAID and add a new function in CATAMID to annotate desmosome).

– This article might assume that readers have some foreground knowledge on both basic anatomy of this animal model and on network theory – not a large community. I thus would appreciate the results to be explained in an intuitive way. For an example, this animal model's overall anatomy should be introduced foremost in the main text. Figure 1A was a nice illustration, but with no explanation in the results, and the single line of figure legend does not offer any information either (I searched online to find some fascinating details in the differences between sg0, sg2, and sg3 that may be relevant to results in later sections).

Including such details will allow readers to put those graphs in their anatomic context naturally, instead of having a reader (e.g. me) need to research each anatomic detail as the name popped up in the text.

– Similarly, all figure legends are exceptionally brief and generally do not provide sufficient guides for the non-expert to understand the results. For example, many figures use CATMAID's output. I believe that circles in most graphs represent the soma of cells, hence the large vs small circles (e.g. in Figure 1E) reflect their size difference. However, this was not explained anywhere, and a general audience would not be able to deduce such information.

2) Below are a few technical questions and discrepancies that authors should address:

– The authors refer to their connectivity map as the 'desmosome connectome'. I think both desmosome and hemidesmosome are included in the final dataset. If so, this needs to be clarified in the sentence where they first defined the term 'desmosome connectome'.

– Page 4: 'The largest cluster of interconnected nodes consists of 2,506 nodes.' I find this confusing. If the total number of nodes is 2,524, this sentence simply means that almost all cells are inter-connected by desmosomes. Is this so? If so, this is not reflected in Figure 1E.

– Page 6: 'The largest Eigenvalue of a graph is another key network property influencing dynamic processes (Restrepo et al. 2006). This value is also largest for the desmosomal graph.'

The authors should provide some context to this and all other graph theory terminology in this section; 'dynamic processes', 'radius of network' etc. appear vague and abstract wordings for a few simple operations that could be explained plainly and interpretated intuitively. E.g. For the second part of the example sentence, I think the authors meant 'The largest Eigenvalue of weighted (?) adjacency matrix of the desmosomal connectivity map'?

– Page 7: 'Nodes corresponding to the neuropodial and notopodial parts of the parapodia also occupy distinct domains (Figure 5A-F).' In Figure 5D, at least sg0 cells appear more much distributed than their physical locations (5C)? The authors should provide more details to explain what is shown in figures.

– Page 4-6: Section 'Local connectivity and modular structure of the desmosomal connectome'. The authors applied an improved clustering algorithm (Leiden) to assign nodes of the biological desmosomal connectome, but later used the Louvain algorithm to compare the modularity scores for their subsampled connectomes with that random networks. The same algorithm should be used to assess and compare community structures.

Reviewer #3 (Recommendations for the authors):

This paper is based on digital reconstruction of a serial EM stack of a larva of the annelid Platynereis and presents a complete 3D map of all desmosomes between somatic muscle cells and their attachment partners, including muscle cells, glia, ciliary band cells, epidermal cells and specialized epidermal cells that anchor cuticular chaetae (circumchaetal cells) and aciculae (circumacicular cells). The rationale is that the spatial patterning of desmosomes determines the direction of forces exerted by muscular contraction on the body wall and its appendages will determine movement of these structures, which in turn results in propulsion of the body as part of specific behaviors.

To go a step further, if connecting this desmosome connectome with the (previously published) synaptic connectome, one may gain insight into how a specific spatio-temporal pattern of motor neuron activity will lead, via a resulting pattern of forces caused by muscles, to a specific behavior. In the authors' words: "By combining desmosomal and synaptic connectomes we can infer the impact of motoneuron activation on tissue movements". This is an interesting idea which has the potential to make progress towards understanding in a "holistic" way how a complex neural circuitry controls an equally complex behavior. The analysis of the EM data appears solid; the authors can show convincingly that desmosomes can be resolved in their EM dataset; and the technology used to plot and analyze the data is clearly up to the task. My main concern is with the way in which the desmosome pattern is entered in the analysis, which I think makes it very difficult to extract enough relevant information from the analysis that would reach the stated goal.

1. The context of how different structures of the Platynereis larval body, by changing their position, move the body needs much more introduction than the short paragraph given at the end of the Introduction.

– My understanding is that the larval body is segmented, and contraction of the segments can cause a certain type crawling or swimming: does it? Do the longitudinal muscles, for example, insert at segment boundaries, and alternating contraction left-right cause some sort of "wiggling" or peristalsis?

– In addition, there are segmental processes (parapodia, neuropodia), and embedded in them are long chitinous hairs (Chaetae, Acicula). Do certain types of the muscles described in the study insert at the base of the parapodia/neuropodia (coming from different angles), such that contraction would move the entire process, including the chaetae/acicula embedded in their tips?

– Or is it that only these chaetae/acicula move, by means of muscles inserting at their base (the latter is clearly part of the story)? Or does both happen at the same time: parapodium moves relative to the trunk, and chaeta/acicula moves relative to the parapodium? How would these movements lead to different kind of behaviors?

– Diagrams should be provided that shed light on these issues.

2. The main problem I have with the analysis is the way a muscle cell is treated, namely as a "one dimensional" node, rather than a vector.

– In the current state of the analysis, the authors have mapped all desmosomes of a given muscle cell to its attached "target" cell. But how is that helpful? The principal way a muscle cell acts is by contracting, thereby pulling the cells it attaches to at its two end closer together. As the authors state (p.4) "…desmosomes..are enriched at the ends of muscle cells indicating that these adhesive structures transmit force upon muscle-cell contraction."

– for that reason, the desmosomes at the muscle tips have to be treated as (2) special sets. Aside from these tip desmosomes there are other desmosomes (inbetween muscles, for example), but they (I would presume) have a very different function; maybe to coordinate muscle fiber contraction? Augment the force caused by contraction?

– As far as I understand for (all of) the desmosome connectome plots, there is no differentiation made between desmosome subsets located at different positions within the muscle fiber. I therefore don't see how the plots are helpful to shed light on how the multiplicity of muscles represented in the graphs cause specific types of neurons.

– As it stands these plots "merely" help to classify muscles, based on their position and what cell type they target: but that (certainly useful) map could have probably also be achieved by light microscopic analysis.

3. Section "Local connectivity and modular structure of the desmosomal connectome" p.4-7" undertakes an analysis of the structure of the desmosome network, comparing it with other networks.

– What is the rationale here? How do the conclusions help to understand how the spatial pattern of muscles and their contraction move the body?

– Isn't, on the one hand (given that position of the desmosome was apparently not considered), the finding that desmosome networks stand out (from random networks) by their high level of connectivity ("with all cells only connecting to cells in their immediate neighbourhood forming local cliques") completely expected?

– On the other hand, does this reflect the reality, given that (many?) muscle cells are quite long, connecting for example the anterior border of a segment with the posterior border.

4. In the section "Acicular movements and the unit muscle contractions that drive them" the authors record movement of the acicula and correlate it with activity (Ca imaging) of specific muscle types. This study gives insightful data, and could be extended to all movements of the larva.

– The fact that a certain muscle is active when the acicula moves in a certain direction can be explained (in part) by the "connectivity": as shown in Figure 7L, the muscle inserts at a circumacicular cell on the one side, and to an epithelial (epidermal?) cell and the basal lamina on the other side. But how meaningful is a description at this "cell type level" of resolution? The direction of acicula deflection depends on where (relative to the acicula base) the epithelial cell (or point in the basal lamina) is located. This information is not given in the part of the connectome network shown in Figure 7L, or any of the other graphs.

https://doi.org/10.7554/eLife.71231.sa1

Author response

Reviewer #1 (Recommendations for the authors):

– It would be interesting to know if muscle cells are not only physically connected by desmosomes but also electrically by gap junctions. It would be interesting if the authors could add data/information about this point to the manuscript.

We have not observed gap junctions in our dataset. We updated the Methods section to mention the absence of observed gap junctions.

– I find the possible acicular movements ('abduction,…'p.8, last paragraph and Figure 6-supp3) very difficult to follow. One problem might be that in Figure 6-supp3, the body axes are mostly missing. In addition, it is inherently difficult to understand spatial and temporal changes using a two-dimensional schematic. If possible, a simulation/video showing how the respective muscle groups lead to different movements of the acicula would be highly informative.

We have added the body axes to this figure (Figure6—figure supplement 3) and simplified it to only show those movements that we could infer confidently. We also included a 3D animation, which shows different muscle groups from different angles (Video 2). This should help to see the different muscle groups relative to the aciculae in the 3D volume.

– Related to the previous point, I have difficulties to follow the relative 4-dimensional movements of the aciculae described in Figure 7G, and a 4D animation (if possible) of the relative movements of the aciculae would certainly help the reader. Alternatively, annotating Video 4 might help understanding acicular movements better.

We have annotated Video 4 to show which are the notopodial and the neuropodial aciculae in the different segments. Our live imaging data did not allow us to track the 4-dimensional movements of the aciculae because we have limited information along the Z-axis. We could tell which chaetae and aciculae are more ventral based on the focus. Tracking the 4D movement of the aciculae would require fast volumetric imaging that we were not able to do in the scope of this project.

Reviewer #2 (Recommendations for the authors):

1) This paper is rich in data and anatomical details, and would be of benefit to scientists who primarily work on other model systems. An obstacle to this goal is that the current draft is dense, making it a slow read (I learned a lot, after several goes). Reorganization can improve this article's readability for a broader audience:

– Some jargon-like technical details distract readers from focusing on main ideas. I would relegate them to Methods (Example: how authors use CATMAID and add a new function in CATAMID to annotate desmosome).

We thank the reviewer for these comments. We have moved some of the technical details to the methods section. We have also shortened the Introduction and simplified the text describing the statistics of the desmosomal connectome.

– This article might assume that readers have some foreground knowledge on both basic anatomy of this animal model and on network theory – not a large community. I thus would appreciate the results to be explained in an intuitive way. For an example, this animal model's overall anatomy should be introduced foremost in the main text. Figure 1A was a nice illustration, but with no explanation in the results, and the single line of figure legend does not offer any information either (I searched online to find some fascinating details in the differences between sg0, sg2, and sg3 that may be relevant to results in later sections).

Including such details will allow readers to put those graphs in their anatomic context naturally, instead of having a reader (e.g. me) need to research each anatomic detail as the name popped up in the text.

We have added more detail about the anatomy of the larva to the Introduction:

“In Platynereis, muscle development starts in the planktonic, ciliated trochophore larval stage (one to two days old) (Fischer et al., 2010). The older (around three to six days) nectochaete larvae have three main trunk segments, each with a pair of appendages called parapodia. The parapodia are composed of a ventral lobe (neuropodium) and a dorsal lobe (notopodium) and each lobe contains a single acicula and a bundle of chitin bristles (chaetae)(Hausen, 2005). Platynereis larvae have an additional cryptic segment between the head and the main trunk segments (Steinmetz et al., 2011). This cryptic segment (also referred to as segment 0) lacks parapodia and gives rise to the first pair of tentacular cirri that start to grow around three days of development.”

We also changed the colouring of the muscles in Figure 1B to highlight the segments more clearly.

– Similarly, all figure legends are exceptionally brief and generally do not provide sufficient guides for the non-expert to understand the results. For example, many figures use CATMAID's output. I believe that circles in most graphs represent the soma of cells, hence the large vs small circles (e.g. in Figure 1E) reflect their size difference. However, this was not explained anywhere, and a general audience would not be able to deduce such information.

We have extended the figure legends including an explanation of what smaller vs larger spheres represent in the morphological renderings.

“Morphological rendering of traced skeletons and soma position (spheres) of all acicular muscles and their classification.”

“ Position and size of cell nuclei, represented as spheres, of all the cells…”.

2) Below are a few technical questions and discrepancies that authors should address:

– The authors refer to their connectivity map as the 'desmosome connectome'. I think both desmosome and hemidesmosome are included in the final dataset. If so, this needs to be clarified in the sentence where they first defined the term 'desmosome connectome'.

We have clarified the definition of the desmosomal connectome.

“These desmosomal connections form a body-wide network through which tensile forces propagate. This network comprises all muscles and all their cellular and non-cellular (e.g., basal lamina) partners. We refer to this network as the desmosomal connectome, with muscles, other cells and basal lamina chunks as nodes and hemidesmosomes and desmosomes as links.”

– Page 4: 'The largest cluster of interconnected nodes consists of 2,506 nodes.' I find this confusing. If the total number of nodes is 2,524, this sentence simply means that almost all cells are inter-connected by desmosomes. Is this so? If so, this is not reflected in Figure 1E.

Almost all cells which have desmosomes are interconnected into a single graph either directly or indirectly through the basal lamina. We clarified this in the text. We also updated Figure 1E (now 1F) to show the edges more clearly.

– Page 6: 'The largest Eigenvalue of a graph is another key network property influencing dynamic processes (Restrepo et al. 2006). This value is also largest for the desmosomal graph.'

The authors should provide some context to this and all other graph theory terminology in this section; 'dynamic processes', 'radius of network' etc. appear vague and abstract wordings for a few simple operations that could be explained plainly and interpretated intuitively. E.g. For the second part of the example sentence, I think the authors meant 'The largest Eigenvalue of weighted (?) adjacency matrix of the desmosomal connectivity map'?

We have thoroughly revised this section, see also our responses to Reviewer #1. We simplified the terminology and discuss fewer network indicators that best highlight the differences between the different types of graphs.

– Page 7: 'Nodes corresponding to the neuropodial and notopodial parts of the parapodia also occupy distinct domains (Figure 5A-F).' In Figure 5D, at least sg0 cells appear more much distributed than their physical locations (5C)? The authors should provide more details to explain what is shown in figures.

We now describe in more detail some cells that are more distributed in the graph than their physical locations.

“Exceptions include the dorsolateral longitudinal muscles (MUSlongD) in segment-0. These cells connect to dorsal epidermal cells that also form desmosomes with segment-1 and segment-2 MUSlongD cells. These connections pull the MUSlongD_sg0 cells down to segment-2 in the force-field layout (Figure 5D).”

– Page 4-6: Section 'Local connectivity and modular structure of the desmosomal connectome'. The authors applied an improved clustering algorithm (Leiden) to assign nodes of the biological desmosomal connectome, but later used the Louvain algorithm to compare the modularity scores for their subsampled connectomes with that random networks. The same algorithm should be used to assess and compare community structures.

We originally used the Louvain algorithm because the Leiden method is computationally ~100 times more intensive. We have now rerun the computations with the Leiden method and updated the relevant code and data. The desmosomal connectome has the highest modularity value also when calculated with the Leiden algorithm, therefore our original conclusions did not change.

Reviewer #3 (Recommendations for the authors):

This paper is based on digital reconstruction of a serial EM stack of a larva of the annelid Platynereis and presents a complete 3D map of all desmosomes between somatic muscle cells and their attachment partners, including muscle cells, glia, ciliary band cells, epidermal cells and specialized epidermal cells that anchor cuticular chaetae (circumchaetal cells) and aciculae (circumacicular cells). The rationale is that the spatial patterning of desmosomes determines the direction of forces exerted by muscular contraction on the body wall and its appendages will determine movement of these structures, which in turn results in propulsion of the body as part of specific behaviors.

To go a step further, if connecting this desmosome connectome with the (previously published) synaptic connectome, one may gain insight into how a specific spatio-temporal pattern of motor neuron activity will lead, via a resulting pattern of forces caused by muscles, to a specific behavior. In the authors' words: "By combining desmosomal and synaptic connectomes we can infer the impact of motoneuron activation on tissue movements". This is an interesting idea which has the potential to make progress towards understanding in a "holistic" way how a complex neural circuitry controls an equally complex behavior. The analysis of the EM data appears solid; the authors can show convincingly that desmosomes can be resolved in their EM dataset; and the technology used to plot and analyze the data is clearly up to the task. My main concern is with the way in which the desmosome pattern is entered in the analysis, which I think makes it very difficult to extract enough relevant information from the analysis that would reach the stated goal.

1. The context of how different structures of the Platynereis larval body, by changing their position, move the body needs much more introduction than the short paragraph given at the end of the Introduction.

– My understanding is that the larval body is segmented, and contraction of the segments can cause a certain type crawling or swimming: does it? Do the longitudinal muscles, for example, insert at segment boundaries, and alternating contraction left-right cause some sort of "wiggling" or peristalsis?

Longitudinal muscles do not insert only at segment boundaries, but have desmosomal connections along the entire length of the cell. Individual longitudinal muscle cells can span up to 3 segments. However the cells are staggered in such a way that all longitudinal muscle cells with somas in one segment can collectively cover up to 4 segments. Longitudinal muscles are involved in turning when swimming (Randel et al., 2014). The undulatory trunk movements and parapodial walking movements are due to the contraction of oblique and parapodial muscles. The longitudinal muscles provide support during crawling (via desmosomal links) but it is unlikely that these muscles contract segmentally. Disentangling the distinct contributions of 53 types of muscles during crawling will require further studies.

– In addition, there are segmental processes (parapodia, neuropodia), and embedded in them are long chitinous hairs (Chaetae, Acicula). Do certain types of the muscles described in the study insert at the base of the parapodia/neuropodia (coming from different angles), such that contraction would move the entire process, including the chaetae/acicula embedded in their tips?

Yes, acicular muscles insert at the proximal base of the acicula, and by moving the acicula they move the entire noto-/neuropodia. We have presented the anatomy of all acicular and chaetal muscles types in the figures and videos.

– Or is it that only these chaetae/acicula move, by means of muscles inserting at their base (the latter is clearly part of the story)? Or does both happen at the same time: parapodium moves relative to the trunk, and chaeta/acicula moves relative to the parapodium? How would these movements lead to different kind of behaviors?

– Diagrams should be provided that shed light on these issues.

We have extended Video 2 to show individual muscles and their relation to the aciculae in one of the parapodia. We also clarified this in the text:

“Several acicular muscles attach on one end to the proximal base of the aciculae and on the other end to the paratrochs and epidermal cells. Oblique muscles attach to the basal lamina, epidermal and midline cells at their proximal end, run along the anterior edge of parapodia and attach to epidermal and chaetal follicle cells at their distal tips. Both of these muscle groups are involved in moving the entire parapodium. Acicular muscles move the proximal tips of the aciculae, while oblique muscles move the parapodium by moving the tissue around the chaetae and the aciculae. All acicular movements also correspond to parapodial movements. Chaetae are embedded in the parapodium and therefore move with it, but the chaetal sac muscles can also independently retract the chaetae into the parapodium or protract them and make them fan out.”

2. The main problem I have with the analysis is the way a muscle cell is treated, namely as a "one dimensional" node, rather than a vector.

– In the current state of the analysis, the authors have mapped all desmosomes of a given muscle cell to its attached "target" cell. But how is that helpful? The principal way a muscle cell acts is by contracting, thereby pulling the cells it attaches to at its two end closer together. As the authors state (p.4) "…desmosomes..are enriched at the ends of muscle cells indicating that these adhesive structures transmit force upon muscle-cell contraction."

At the level of the current analysis our data reveal which cells may be moved by the contractions of the individual muscle cells. The reviewer is right that treating a muscle as a vector (or set of vectors) would be a more accurate description, which would potentially also open up the possibility of computational modelling. We have provided such a vectorised dataset in the revised version, where each muscle-cell skeleton is subdivided into short linear segments (Figure2–source–data 2). This dataset may be useful to approach the problem with a three dimensional approach, which is beyond the scope of the current analysis. We also included an additional video (Video 7) showing examples of muscles and their partners where the cells and the desmosomes connecting them are highlighted. This reveals that the desmosomes connecting two cells are often at the very end of the muscle cell.

– for that reason, the desmosomes at the muscle tips have to be treated as (2) special sets. Aside from these tip desmosomes there are other desmosomes (inbetween muscles, for example), but they (I would presume) have a very different function; maybe to coordinate muscle fiber contraction? Augment the force caused by contraction?

Desmosomes between muscles only occur between muscles of different types, not for homotypic connections. There are other types of junctions (adhaerens-like junctions) that connect individual cells of a muscle bundle together (not analysed here). We clarified this in the text.

– As far as I understand for (all of) the desmosome connectome plots, there is no differentiation made between desmosome subsets located at different positions within the muscle fiber. I therefore don't see how the plots are helpful to shed light on how the multiplicity of muscles represented in the graphs cause specific types of neurons.

We would like to point out that the cells and structures that muscles connect to via desmosomes are very likely the parts of the body that will move during the contraction of the muscle or will provide structural support (e.g. basal lamina) for the muscle cell to contract. This is most evident in the parapodial complex. The majority of muscles in the body connect to the aciuclar folliclecells and the aciculae are the most actively moving parts in the body during crawling (see Video 4). In any case, since we provide all skeleton reconstructions and the xyz coordinates of all desmosomes, the data could be further analysed following these suggestions by the reviewer.

– As it stands these plots "merely" help to classify muscles, based on their position and what cell type they target: but that (certainly useful) map could have probably also be achieved by light microscopic analysis.

This has never been achieved by light microscopy analysis in the hundreds of papers on invertebrate muscle anatomy (e.g. by phalloidin staining). For an LM analysis, it would not be sufficient to label the muscle fibres, but one would also need to label the desmosomes and a multitude of non-muscle cell types including the extent of their cytoplasm. This is technically very challenging (we would nevertheless be happy to hear specific suggestions for markers etc. from the Reviewer). Currently, only EM provides the required depth of structural information and resolution. This is why we believe that our dataset and analysis is unique, despite over a century of research in invertebrate anatomy.

3. Section "Local connectivity and modular structure of the desmosomal connectome" p.4-7" undertakes an analysis of the structure of the desmosome network, comparing it with other networks.

– What is the rationale here? How do the conclusions help to understand how the spatial pattern of muscles and their contraction move the body?

We hope that our analysis may also be of interest to the community of network scientists and we believe that the reconstruction of a quite large and novel type of biological network warrants a more quantitative network analysis, using the standard methods and measures of network science – as we presented e.g. in Figure 4 – even if these mathematical analyses may not directly reveal how muscles move the body. We hope that some readers with an interest in quantitative analyses will also appreciate the broader picture here.

– Isn't, on the one hand (given that position of the desmosome was apparently not considered), the finding that desmosome networks stand out (from random networks) by their high level of connectivity ("with all cells only connecting to cells in their immediate neighbourhood forming local cliques") completely expected?

We disagree that the result was completely expected. Even if this was the case, we think it is quite different to say that a result is expected or to thoroughly quantify certain parameters and mathematically characterise key properties of the desmosomal graph (as we have done). These network analyses help to conceptualise our findings and to think about the muscle system in more global, whole-body terms.

– On the other hand, does this reflect the reality, given that (many?) muscle cells are quite long, connecting for example the anterior border of a segment with the posterior border.

Indeed, a quantitative analysis helped us to identify cases where the reality deviated somewhat from what was completely expected, and we thank the reviewer for these comments. As we explain in the revised version, some longitudinal muscles show an unexpected position in the force-field layout of the graph, due to their long-range connections. We have added extra clarifications to the text:

“To analyse how closely the force-field-based layout of the desmosomal connectome reflects anatomy, we coloured the nodes in the graph based on body regions (Figure 5). In the force-field layout, nodes are segregated by body side and body segment. Exceptions include the dorsolateral longitudinal muscles (MUSlongD) in segment-0. These cells connect to dorsal epidermal cells that also form desmosomes with segment-1 and segment-2 MUSlongD cells. These connections pull the MUSlongD_sg0 cells down to segment-2 in the force-field layout (Figure 5D).”

4. In the section "Acicular movements and the unit muscle contractions that drive them" the authors record movement of the acicula and correlate it with activity (Ca imaging) of specific muscle types. This study gives insightful data, and could be extended to all movements of the larva.

– The fact that a certain muscle is active when the acicula moves in a certain direction can be explained (in part) by the "connectivity": as shown in Figure 7L, the muscle inserts at a circumacicular cell on the one side, and to an epithelial (epidermal?) cell and the basal lamina on the other side. But how meaningful is a description at this "cell type level" of resolution? The direction of acicula deflection depends on where (relative to the acicula base) the epithelial cell (or point in the basal lamina) is located. This information is not given in the part of the connectome network shown in Figure 7L, or any of the other graphs.

This information is indeed not shown in the graphs, where each cell is treated as a node. However, we provide this information in the detailed anatomical figures in Figure 6 —figure supplement 1-3 and Video 7, where the individual acicular and oblique muscle types are visualised. In principle, one could subdivide aciculae into e.g. proximal and distal halves and derive a more detailed network. We have not done this but since all the EM, anatomical rendering and connectivity data are available in our public CATMAID server (https://catmaid.jekelylab.ex.ac.uk/), we hope that the interested readers will be able to further analyse the data.

We renamed ‘epithelial’ cells to ‘epidermal’ cells.

https://doi.org/10.7554/eLife.71231.sa2

Article and author information

Author details

  1. Sanja Jasek

    Living Systems Institute, University of Exeter, Exeter, United Kingdom
    Contribution
    Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft, Writing – review and editing
    Competing interests
    No competing interests declared
  2. Csaba Verasztó

    Living Systems Institute, University of Exeter, Exeter, United Kingdom
    Present address
    EPFL Campus Biotech, Geneva, Switzerland
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared
  3. Emelie Brodrick

    Living Systems Institute, University of Exeter, Exeter, United Kingdom
    Contribution
    Data curation, Formal analysis, Investigation, Visualization, Methodology, Writing - original draft, Writing – review and editing
    Competing interests
    No competing interests declared
  4. Réza Shahidi

    Living Systems Institute, University of Exeter, Exeter, United Kingdom
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Writing – review and editing
    Competing interests
    No competing interests declared
  5. Tom Kazimiers

    1. Janelia Research Campus, Ashburn, United States
    2. kazmos GmbH, Dresden, Germany
    Contribution
    Software, Methodology, Writing – review and editing
    Competing interests
    Tom Kazimiers is the founder of kazmos GmbH, a company that continues the development of the open-source package CATMAID
  6. Alexandra Kerbl

    Living Systems Institute, University of Exeter, Exeter, United Kingdom
    Contribution
    Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared
  7. Gáspár Jékely

    Living Systems Institute, University of Exeter, Exeter, United Kingdom
    Contribution
    Conceptualization, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing – review and editing
    For correspondence
    g.jekely@exeter.ac.uk
    Competing interests
    Reviewing editor, eLife
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8496-9836

Funding

European Commission (FP7-PEOPLE-2012-ITN grant no. 317172)

  • Sanja Jasek
  • Gáspár Jékely

Wellcome Trust (Investigator Award 214337/Z/18/Z)

  • Sanja Jasek

European Research Council (grant agreement No 101020792)

  • Alexandra Kerbl
  • Gáspár Jékely

The funders had no role in study design, data collection, and interpretation, or the decision to submit the work for publication. For the purpose of Open Access, the authors have applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission.

Acknowledgements

We thank Liz Williams and members of the Jékely lab for comments on the manuscript. We thank Luis A Bezares-Calderón for providing GCaMP larvae. We also thank Kirsty Wan for helping with the fast DIC imaging of acicular movements. This research was supported by the FP7-PEOPLE-2012-ITN grant no. 317172 'NEPTUNE'. This research was funded by the Wellcome Trust Investigator Award 214337/Z/18/Z. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 101020792).

Senior Editor

  1. Marianne E Bronner, California Institute of Technology, United States

Reviewing Editor

  1. Kristin Tessmar-Raible, University of Vienna, Austria

Publication history

  1. Preprint posted: June 10, 2021 (view preprint)
  2. Received: June 12, 2021
  3. Accepted: December 7, 2022
  4. Accepted Manuscript published: December 20, 2022 (version 1)
  5. Version of Record published: January 25, 2023 (version 2)

Copyright

© 2022, Jasek et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 245
    Page views
  • 41
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Sanja Jasek
  2. Csaba Verasztó
  3. Emelie Brodrick
  4. Réza Shahidi
  5. Tom Kazimiers
  6. Alexandra Kerbl
  7. Gáspár Jékely
(2022)
Desmosomal connectomics of all somatic muscles in an annelid larva
eLife 11:e71231.
https://doi.org/10.7554/eLife.71231
  1. Further reading

Further reading

    1. Cell Biology
    Hiroaki Ishikawa, Jeremy Moore ... Wallace F Marshall
    Research Article Updated

    Eukaryotic cilia and flagella are microtubule-based organelles whose relatively simple shape makes them ideal for investigating the fundamental question of organelle size regulation. Most of the flagellar materials are transported from the cell body via an active transport process called intraflagellar transport (IFT). The rate of IFT entry into flagella, known as IFT injection, has been shown to negatively correlate with flagellar length. However, it remains unknown how the cell measures the length of its flagella and controls IFT injection. One of the most-discussed theoretical models for length sensing to control IFT is the ion-current model, which posits that there is a uniform distribution of Ca2+ channels along the flagellum and that the Ca2+ current from the flagellum into the cell body increases linearly with flagellar length. In this model, the cell uses the Ca2+ current to negatively regulate IFT injection. The recent discovery that IFT entry into flagella is regulated by the phosphorylation of kinesin through a calcium-dependent protein kinase has provided further impetus for the ion-current model. To test this model, we measured and manipulated the levels of Ca2+ inside of Chlamydomonas flagella and quantified IFT injection. Although the concentration of Ca2+ inside of flagella was weakly correlated with the length of flagella, we found that IFT injection was reduced in calcium-deficient flagella, rather than increased as the model predicted, and that variation in IFT injection was uncorrelated with the occurrence of flagellar Ca2+ spikes. Thus, Ca2+ does not appear to function as a negative regulator of IFT injection, hence it cannot form the basis of a stable length control system.

    1. Cell Biology
    2. Microbiology and Infectious Disease
    Yi Fan, Ping Lyu ... Chenchen Zhou
    Tools and Resources

    Oral inflammatory diseases such as apical periodontitis are common bacterial infectious diseases that may affect the periapical alveolar bone tissues. A protective process occurs simultaneously with the inflammatory tissue destruction, in which mesenchymal stem cells (MSCs) play a primary role. However, a systematic and precise description of the cellular and molecular composition of the microenvironment of bone affected by inflammation is lacking. In this study, we created a single cell atlas of cell populations that compose alveolar bone in healthy and inflammatory disease states. We investigated changes in expression frequency and patterns related to apical periodontitis, as well as the interactions between MSCs and immunocytes. Our results highlight an enhanced self-supporting network and osteogenic potential within MSCs during apical periodontitis-associated inflammation. MSCs not only differentiated towards osteoblast lineage cells, but also expressed higher levels of osteogenic related markers, including Sparc and Col1a1. This was confirmed by lineage tracing in transgenic mouse models and human samples from oral inflammatory-related alveolar bone lesions. In summary, the current study provides an in-depth description of the microenvironment of MSCs and immunocytes in both healthy and disease states. We also identified key apical periodontitis-associated MSC subclusters and their biomarkers, which could further our understanding of the protective process and the underlying mechanisms of oral inflammatory-related bone disease. Taken together, these results enhance our understanding of heterogeneity and cellular interactions of alveolar bone cells under pathogenic and inflammatory conditions. We provide these data as a tool for investigators not only to better appreciate the repertoire of progenitors that are stress responsive but importantly to help design new therapeutic targets to restore bone lesions caused by apical periodontitis and other inflammatory-related bone diseases.